Hacker News Reader

Comment by izacus | Hacker News Reader

How did we get from a situation where you could just pay with cash/card for fuel to having monopolized single-brand car fueling stations?

View on HN

When it started taking 30 minutes to "fill up" your car.

High speed chargers can cost >$200k installed.

A gas pump costs about $25k installed.

If the average person spends 4 minutes at a pump, that comes out to:

The time value of a slot at a super charger can be >$5 for each charge.

The time value of a slot at a gas pump is ~$0.07 per fill up.

The economics of a charging business are awful. High CapEx, few cars, not many of them need chargers since they can charge at home.

The economics of a gas station were not terrible.

There's a reason you don't see immigrants from all over the world coming to the US to open charging stations the way you saw them opening gas stations.

And that's the reason you have charging monopolies.

>High speed chargers can cost >$200k installed. >A gas pump costs about $25k installed.

These numbers seem skewed to me. I think you're quoting prices for gas pumps with just the pump, not any storage infrastructure like underground tanks, stores, employees. Then you're quoting the cost high speed chargers without any kind of electrical infrastructure to support them, which would be expensive, but more analogous to including the cost of all the other infrastructure for gas stations as well.

Storage tanks are (relatively) cheap. Stores and employees are needed because margins on gasoline are extremely thin, and it's an easy way to make money if you're already putting in the pumps.

The infrastructure for high speed chargers is a lot of metal, which unlike a store, does not contribute extra money in addition to the pump.

Since fast DC charging takes more time, there's a lot more time for a person to buy something from the store.

Gas stations depend on cycling people through quickly during rush hour. A half dozen people sitting around for half an hour at a time won't spend enough to make up the difference in volume of traffic a normal station gets, and they don't typically devote space for lounges or other amenities unless they're nice long haul truck stops.

The model doesn't have to be like for like though. Set up a restaurant instead of a convenience store (think American diner style) and offer a discount on your charge for every $10 spent, or a similar integration.

Now you have a reason for someone to sit and spend money, and likely enough as EVs hit critical mass to justify costs. There are things to work out on speed of food vs charging times, potentially a queue and valet setup - but napkin maths it seems viable.

Tanks are cheap. Land for them, and digging, are not.

Employees are needed because most places require gas stations to be manned. Unless you're talking Bucee's they're not making much from selling a few drinks, they're just making marginal use of the building.

Gas stations make very little money from selling gas. The vast majority of their profits come from the building itself, to the point that if they only sold gas, they wouldn't be able to operate in most cities.

I would think that given the chargers are so expensive, they'd want to make them easy to pay to use.

I have a PHEV, so using a charger out and about is silly, but I've only managed to use one pay charger, and it was a lot harder than a gas pump. Even the free ones are hard to use. With the PHEV, the charged amounts are rather low, so I understand the desire to push stored value accounts, but really there should have been colaboration and portability, and at least chargers should have clear information on how to make them start. Go to some website and here's my name on the website. I've seen some that do, but many that don't.

Gas pumps aren't always easy either. I've had several cars where you have to hold it right or the pump stops, but I've never had much trouble giving them money. Also had fun when VW sent out anti-misfueling inserts for my TDI and I got to learn which diesel pumps had unleaded gasoline sized nozzles and which had leaded gasoline sized nozzles; before the insert, the car would take either, after only the leaded fuel size.

Charging via application brings unnecessary friction and severely degrades UX.

Another UX offenders are QR codes on charging stalls and locking charging cables unless you will find that correct QR code so you can unlock your charging cable. Not that one, that will unlock charging cable in the stall next to you. For FFS why do I need to search for QR codes at a first place...

I can imagine a simple solution - take a cable, connect it to the car and start changing by tapping a contactless credit card reader next to your cable. But I never saw simple solutions, seems like every provider wants to torture its customers with half assed applications.

You can do this in france. Tap once to open the hatch for the port, tap when you're done to pay for what you've used.

You don't even need the contactless credit card reader. ISO 15118 is the standard for plug and charge.[1] Tesla implements it, as do most other EV manufacturers. The problem is that charging networks and manufacturers haven't settled on a way to exchange and update certificates.[2]

1. https://en.wikipedia.org/wiki/ISO_15118

2. https://www.theverge.com/2024/12/4/24312002/ev-plug-and-char...

Tesla doesn't use the ISO standard for plug&charge, they have their own custom implementation.

Before the ISO standard existed, Tesla had their own implementation of plug and charge. They've since added support for ISO 15118, allowing non-Teslas to charge at supercharging stations without using the Tesla app.[1]

The issue is that the ISO standard relies on TLS certificates, but manufacturers and charging networks have not yet agreed upon a standard set of CAs. Tesla wants to update the standard to remove the TLS requirement, which would improve reliability and time to start charging. But signed metering receipts are broken in the existing standard, so that needs to be fixed before the TLS requirement can be removed.

1. See page 3 of https://efiling.energy.ca.gov/GetDocument.aspx?tn=256283

Thank you! That's pretty cool. The proposed changes actually will fix this glaring omission:

> Limited Security – by only validating SECC TLS cert is from a trusted issuer, one charger’s compromised private key compromises the entire region

And I like the simplification. Instead of relying on validating contracts, the charger provider will simply rely on signed "metering receipts" from the car. Each car has its own private key (presumably in some hardware-hardened storage), and the charging network can just associate the payment details with the public key of the car.

The provider can use the receipts as a proof that the car has indeed used the charging equipment. And the receipts are sent periodically during the charging process, so the charger can terminate the session if there's a discrepancy between the station's and the car's accounting.

Nice and neat.

Edit: and this also can easily work offline. The networks can just sync the list of approved public keys to chargers with the corresponding credit balances. It'll require account setup with each network, but if you have to do it once, it's not _too_ bad.

FWIW, if the manufacturer does it right, it could theoretically do the sync once with your car and have it authenticate with a lot of different networks that participate in that same partnership. Otherwise, it'll just be a one-time setup the first time you charge.

For example, Ford has this "Blue Oval" network concept, so any charger network that is a part of that would trust that without necessarily needing me to associate my individual car identity.

Honestly though I'm kind of a fan of just having a credit card reader on the dispenser. Its way easier if I want to choose a different payment method for a particular charge, and honestly it is not that much additional work to plug in, tap a credit card or phone, and then it starts charging. Its adding like 10-30 seconds to a 10min+ transaction.

Oh, for sure. There's a lot of possible workflows. E.g. a car can present the provider's login screen on the dashboard.

> Its way easier if I want to choose a different payment method for a particular charge, and honestly it is not that much additional work to plug in, tap a credit card or phone, and then it starts charging. Its adding like 10-30 seconds to a 10min+ transaction.

Credit card readers are a PITA, and they need connectivity. More importantly, the ISO 15118 protocol can be used with wireless charging! Imagine just parking at a designated spot, clicking "confirm" on the car dash, and walking away. The car can even align itself with the charging coils.

> Credit card readers are a PITA, and they need connectivity.

EVV transactions can be done offline. It depends on the issuer if it's allowed.

They do on newer chargers.

After my fair share of car being incompatible with basic CCS (Connector fits, car tells me it does not like the electricity here) I just want the charging to be working without some fancy features which charger manufacturers are going to implement wrong anyway.

I'd really like to know which cars and which networks have been the issues you've experienced.

I've seen people in BMW's and VW's pull up to an EA charger, try getting it to charge a few times, and have it fail every time. Then I see a Hyundai or a Ford pull up, and it works without issue.

I agree. ChargePoint has the contactless payment setup, but you need to have the Application setup with your account on it. After that it works through the Apple Wallet app.

Add on to that that many gas stations make more money from going in and buying something in the store. If you stay at an EV charging station for longer, you are going to be more likely to buy something. This is a reason there are Tesla superchargers now at places like sonic in the west. Also see Buckees.

I imagine battery technology improving fast enough that it stays economically challenging to build out charging to anything like gas station level scale.

We will come to think of stopping to fill up your car routinely as quaint as lamp lighters.

This doesn’t seem right to me. Is this 200k per dispenser? The dispenser is really just a fancy switch and a plug in a kiosk. If you are talking about the central transformer/switching systems, then yes that makes sense. But you can add a lot of dispensers to that.

A pump is only 25k to install if you don’t include the infrastructure to support the pump (tank, canopy, fire suppression, filters, etc).all that costs more than 200k.

Let’s say 25k is the marginal cost for an extra pump. What is the marginal cost for an extra dispenser?

Exactly. I would be surprised if frankly Tesla's cost to open a decent supercharging station (6-10 stalls or something) is greater than the cost to open a 4 stall gas station.

DCFCs are much more than a “fancy switch”. It’s a circuit capable of converting 3-phase high voltage AC into variable high voltage DC at 150-350kW. The power electronics are very, very expensive. As far as I’m aware, the large transformer you see usually hidden somewhere nearby is not the primary cost (although it is expensive, especially for very high wattage ones).

And yes, it’s 200k per dispenser including the infrastructure. It doesn’t scale as well as you think it does. I think Tesla has been quoted as around 50k per dispenser including infra though, so some of it is just poor efficiency in costs by other mfgs.

The fast charger is the expensive part, the dispenser is not nearly so. Just like a gas station, dispensers to chargers are many-one. I was being a bit glib when I said the dispenser is a fancy switch (esp if the lines are cooled) but only just a bit.

I see a report that has Tesla’s cost as 43k per installed dispenser. That is a fully load cost, not the marginal cost of dispenser but it is good enough.

Looking at listings for gas stations for sale (with a convenience store but no auto repair), I see about 150-300k per dispenser. That isn’t exactly apples to apples but suffice to say it isn’t exactly cheap and much closer to representing the cost than the cost of a pump (which is I assume cheaper than a dispenser).

At least the Tesla implementation of their latest stuff, dispenser to faster charger ratio is 1:1. Their older design would gang 2 dispensers to a single 150kW combined charge rate, but new ones are 250kW per dispenser for each port simultaneously.

The expensive power electronics you speak of are not part of the dispenser. The dispenser is really just a very fancy switch. It performs some payment authorization and then switches on those expensive power electronics.

If you look at a modern Electrify America unit, the dispenser is an extremely slim panel with a screen. It clearly isn't big enough to contain these power electronics.

Now transformer isn't a great name because it implies an AC-to-AC device which this is not. So I can see where the confusion comes from.

I’m not confused here, to be clear. There is a giant transformer that converts from mid voltage AC to low voltage AC at most installations. Because they draw so much power they need their own dedicated transformers. There’s sometimes (depending on the installation) a separate cabinet to convert the AC to DC and then transfers it to dispensers, though, as you say. The more modern high power ones tend to use separate cabinets, I’d agree. Older ones could take 3-phase AC directly since they’re much lower power overall.

Those older ones were just normal step-down transformers. In Europe that would often be three-phase, in the US it would normally be two-phase (J-1772 is only two-phase compatible). In that world, the "charger" (EVSE) was just a smart relay which would tell the car how much power the circuit would handle, but the actual charger was a thing in the car. That's commonly called Level 2 charging.

What is commonly talked about here are DC fast chargers, where the actual "charger" is an AC to DC cabinet on-site. Those chargers could be connected to one or many dispensers. The dispensers are the things with the cables that plug into the car, handle payment negotiation, and relay battery state to the charger in the cabinet. You could have multiple dispensers to chargers in this setup.

The technical English parlance here would be EVSE for the Level 2 chargers (the smart relay things), and dispensers for the pedestals with the cables for DC fast chargers (DCFCs).

Again, not confused here. I’m talking about older stuff like the 50kW CCS/chademo systems. There’s no “older” with standard level 2 charging. Level 2 charging coexists with level 3 and will continue to coexist.

That's because the power electrics are in a big white cabinet on the side.

DCFC electrics are absolutely insane.

> The time value of a slot at a gas pump is ~$0.07 per fill up

I think charging stations have an asset that you have neglected to consider: a captive audience.

Yeah, there will always be people who plug in their car and then scroll for half an hour. But there are also people who would be interested in grabbing a bite to eat, walking around, spending money on something dumb, etc. Having a charging station be a place that you can spend money on human amenities in addition to charging improves the economics.

But instead our charging stations are like, three plugs in the back of a hotel parking lot.

> But instead our charging stations are like, three plugs in the back of a hotel parking lot.

It's almost as if there's an economic reason for that.

Maybe it's a chicken-and-egg situation that needs a tipping point. If there's enough EVs on the road it becomes profitable.

Companies tend not to publish their costs, but we can find some info about the economics of charging stations. A grant program in Texas subsidized up to 70% of the cost of charging stations. Tesla bid $500k for a fast charging station with 17 plugs, and around $380k for a station with 9 plugs.[1] That's $29k per plug for the 17 charger station. Assuming they asked for the maximum subsidy, that means Tesla's costs are around $41k per plug.

If Tesla bills customers 20 cents per kWh above market electricity rates, they'll make back the cost of the plug after 205,000kWh. A typical EV has a 70-100kWh battery, so that's 2,000-3,000 full charges. If it takes an hour to fully charge a car, then each plug will be profitable after 2-4 months assuming a 100% duty cycle. Actual duty cycles are much lower. If we assume the charger is active 15% of the time, then it will be profitable after 18-30 months. Again, these are all rather pessimistic estimates. Actual markups and duty cycles are higher, and actual charging times are shorter. If we assume a 20% duty cycle, 45 minute charging time, and a 30 cent per kWh markup, it only takes 8-10 months to become profitable. I didn't include maintenance costs, but those are much lower than gas stations because charging stations require no full time staff.

1. https://www.forbes.com/sites/bradtempleton/2022/04/14/teslas...

EV manufacturers (re: most car manufactures) are going to do all they can to get the government to subsidize this as much as possible.

I noticed an explosion of EV charging stations where I live after the state government put incentives in to subsidize their build out. Other states have done it already and are doing it now.

They're going to see the pattern of "I can get the government to subsidize this" and will hold out for them going forward, is my best guess.

It's important to distinguish Tesla from other companies. The others haven't figured out how to make chargers cheaply yet, so they're forced to ask for subsidies. That's why Tesla's bids on that Texas contract were so much lower than the competition.

Tesla mass produces their charging stations in factories. Entire stations are assembled from modular units, mounted on a concrete slab, and shipped to the installation site on a flatbed.[1] These optimizations have gotten installation times down to 4 days.[2] Sources inside of Tesla have claimed their charging network is profitable, which is impressive considering that until recently, their network got no subsidies because it only supported charging Teslas.

1. https://x.com/Tesla/status/1636085524694695939

2. https://x.com/TeslaCharging/status/1777395990766432596

The others don't _want_ to make chargers. They don't or want to know how.

Charging infrastructure is a public good so it ought to be the government's job to incentivize it, but Tesla's supercharger network works just as well.

Personally, I don't want my car company to also be the company I buy the fuel from. I'd want a marketplace of people competing to sell me fuel.

I don't want to only be able to fuel my Hyundai at Hyundai-affiliated gas stations. And it'd be bullshit if I couldn't get gas from the Ford station because I'm in a Hyundai.

I get Tesla needed to jumpstart the industry, but if its going to actually work the charging industry needs to stand on its own and not rely on the automakers making their own networks.

Tesla has yet to ever produce a 1000V charging cabinet. Their v4 chargers that claim to be 1000V aren’t, because they are still driven by 400V cabinets.

Building 1000V cabinets is very very expensive, which is why Tesla is behind the curve and hasn’t done it yet.

EA cabinets are all 1000V. You are comparing apples and oranges.

And 400V is the past.

Is it? This makes no real difference to my entire point, which is companies are learning that they should wait for subsidies, and will either build out much slower or not at all until they do. State governments have now set precedent for it.

It’s not even that I’m against subsidies, to be clear, but they’re becoming pro forma and they aren’t terribly well thought out on setting goals.

So they get subsidized to be built where they were going to build the chargers anyway

> The economics of a charging business are awful.

Ish. I mean one thing the charging business has going for it is that after the initial investment the ongoing investment is basically 0. You don't have weekly fuel shipments or any real ongoing maintenance costs.

You also don't see immigrants all over charging stations because they are mostly unmanned. Usually placed in parking lots of retail stores.

> I mean one thing the charging business has going for it is that after the initial investment the ongoing investment is basically 0

That's not actually correct. You need to pay for reserved power in the grid, which can be significant amount regardless if you are using that power or not. So installing big DCFC charger where nobody is going to use it will eat you up on fees for reserved power.

Any insights to what the reserved rates are? Nothing really came up on google. I could find some commercial power rates which tended to be lower than residential rates, but I assumed that's not accurate since these things will eat a lot more power than most businesses would.

> You also don't see immigrants all over charging stations because they are mostly unmanned. Usually placed in parking lots of retail stores

tf is this supposed to mean

Mostly just responding to the OP

> There's a reason you don't see immigrants from all over the world coming to the US to open charging stations the way you saw them opening gas stations.

My guess is what they meant by this is the very visible minimarts on gas stations often manned by (presumed) immigrants.

> There's a reason you don't see immigrants from all over the world coming to the US to open charging stations the way you saw them opening gas stations.

As the other commenter said, AFAIK, the majority of a gas station's income comes from the convenience store, no? Profit margins on gas are really low - Google says it's around 2%.

I imagine there are additional factors.

* many prime charging station locations are where the gas stations currently are. The high installation costs and long charging times mean you’d lose money swapping them 1:1 right now, and most gas stations are already optimized to hold the max pumps per spot of land.

* it is challenging to find new locations not least because cities have found out that gas stations are pretty problematic uses of land when it comes to neighborhood desirability, traffic, etc. so many are loth to approve new charging locations.

The cons of a gas station in the neighborhood are when it's a convenient store with pumps. The sale of lotto, beer, smokes, and low cost food attract it's own crowd that the NIMBYs do not like. Switch that to a storeless unmanned set of pumps, and all of that goes away. I've seen a few of these types of locations around town. Just an awning with a few pumps under it. That's pretty much what all of the electric charge locations look like. Just "rent" spots at existing places on the highway with parking nearest the main road (which is typically the least desired spots in the lot). Pretty much any hotel would be a good candidate.

Eh there are a ton of other downsides. By their nature, high-throughput charging stations with lots of curb cuts are bad for walkability, eyes on the street, etc. that people want in their neighborhoods. For example many cities now ban new drive throughs.

And the convenience stores are the thing that makes the money at gas stations.

> And the convenience stores are the thing that makes the money at gas stations.

Because the gas is expensive, you need to hold inventory (and have a tank that needs to be insured and inspected and ...) and a lot of the profit is going to huge oil companies.

If you are running chargers, there is no inventory and the cost is whatever the electricty costs are.

I suspect that chargers could actually be a profit center as opposed to gas pumps which barely break even and sometimes go negative in profits.

EV stations have a lot more competition that doesn’t need to turn a profit; a person’s home. Nobody has home gas tanks to fill up their car.

I don’t know many people buying electric without having a home charger.

Literally anyone who lives in an apartment.

That's a pretty big exaggeration.

It's not easy to install your own in an apartment, but increasing numbers of apartments have chargers and most apartments could have them.

Sure, they are slowly getting added, but the incentives are pretty low. The vast majority of apartment-dwellers charge either at work (if they are lucky enough to work for a large enough company that builds chargers) or at a public charging station

that's the thing, i have a lot of friends who are in apartments, and know very few of them going fully EV

A gas pump relies on a significantly larger chain of infrastructure though. We just accept that it's quicker because it's already in place. No one factors in the time at the refinery etc etc.

Energy delivered in the from electricity isn't that bad or slow when you compare it fully, including the generation infrastructure.

I personally think we will move towards swappable pre-charged batteries at charging stations, which will drop the "refuel" time to something comparable to (or less than) petrol.

> High speed chargers can cost >$200k installed.

Should this go down as chargers become more popular due to economies of scale? Or will they always be significantly more expensive than gas pumps?

> from a situation where you could just pay with cash/card for fuel to having monopolized single-brand car fueling stations?

The Model T was released in 1908 [1]. Standard Oil was broken up in 1911 [2], 6 years after Bowser added a hose attachment to his pumps [3].

[1] https://en.m.wikipedia.org/wiki/Ford_Model_T

[2] https://en.m.wikipedia.org/wiki/Standard_Oil_Co._of_New_Jers...

[3] https://www.saferack.com/evolution-gas-pump/

Lack of ambition. Regulation solves the problem. Europe has mandated contactless payment for all chargers.

Here's a Tesla V4 charger charging a Kia with contactless payment, no Tesla account and no Tesla app:

https://www.youtube.com/shorts/yflZN0dLT8s

Because you can't DRM a liquid but you can DRM a charger. Vertical monopolies are always a good business plan when there's no regulations to prevent you from forcing people to buy your overpriced "premium certified" garbage instead of paying market rate for the same thing from a generic brand.

No laws on campaign financing. You let money rule democracy. And now the CEO of a car company is part of the US government.

The last several times I did a DC fast charge that wasn't just plug and charge it had a credit card terminal that took chipped cards, swiped cards, and tap to pay. Plug in, tap credit card, it starts charging.

All the chargers I use regularly (Chargepoint, EVgo) support tap to pay with cards. No app or account needed.

Most new chargers now have chip/tap readers. Including Tesla's V4 chargers.

Its the same story that gasoline went through

but yeah the current situation is currently bad, let me say something louder for the entrepreneurs in the back: undermaintained last resort fast-charging is one thing, but if you have a slow-charging network I'm not downloading your app

Reply on news.ycombinator.com