EV Charging Statistics US 2026: 312,486 Public Ports, 186,420 Sessions and the Complete National Picture

The 2026 headline findings#

The United States had 312,486 public electric vehicle charging ports at 89,412 stations on 30 June 2026. EV Cable Hub's 2026 census recorded growth of 26.4% on the previous year, with 65,266 ports added in twelve months, and 27.0% of the national total is now DC fast.

Two counting units run through this dataset and they are not interchangeable. A port is a single plug a driver can use. A station is the site those ports stand on, and it carries 3.50 ports on average. This page reports ports as its headline unit throughout, because that is the number that decides whether a driver arriving at a site can plug in. Station counts are the unit most often quoted in political coverage, which is the single largest source of confusion in US charging reporting.

Five structural facts run underneath everything else here. The network is growing at roughly 5,400 ports a month. It is heavily concentrated in a handful of states, with California alone holding 19.5% of the national total. It remains majority Level 2, at 71.8% of ports. Its reliability is materially worse than the comparable European figure, with 7.4% of attempted sessions failing at the first attempt. And the overwhelming majority of actual charging energy never touches it, because 79.2% of it is delivered at home.

Those five facts interact in ways that make single-number summaries misleading. A network that is majority Level 2 is not failing if those ports are doing overnight and workplace duty. A national provision figure of 91 ports per 100,000 residents says nothing useful to a driver in one of the 284 counties with no DC fast port at all. The sections that follow take each apart: counting conventions first, then growth, charging levels, networks, states, metros, deserts, NEVI, connectors, home charging, sessions, cost, reliability and the 2030 outlook.

How many public charging ports the US has, and how they are counted#

The US has 312,486 public charging ports at 89,412 stations as of 30 June 2026, an average of 3.50 ports per station. The two numbers describe the same network measured two ways, and the gap between them is why published US charging totals disagree by a factor of three.

This is the credibility section and it comes second on purpose. Port-versus-station confusion is the single most common error in US charging coverage, and correcting it publicly is worth more than any other paragraph on this page. A headline announcing a hundred thousand chargers and a headline announcing three hundred thousand can both be accurate descriptions of the same network in the same month.

EV Cable Hub's 2026 census counts publicly accessible charging ports in all fifty states and the District of Columbia, whether free or paid and whether available around the clock or during restricted hours only. It excludes residential units, workplace ports behind a gate with no public access, ports out of service for more than 90 days, dealer-only ports and private fleet depots. Those exclusions are large: workplace and fleet ports alone account for a further 166,812 ports that exist and work but that no member of the public can drive up to and use.

Three counting errors circulate widely and each has a figure attached. The first is quoting stations as though they were ports, which understates the plugs a driver can use by 223,074. The second is including announced or awarded but not yet energised ports, which adds capacity that cannot charge a car today. The third is counting private fleet depots inside a public total. Read the reconciliation table below and any published US figure can be mapped back to this one.

One further convention affects every state and metro figure on this page. Ports are attributed to the county and state the site physically sits in, not to the network's own regional structure and not to a corporate billing address. That distinction sounds academic until you reach a station on a state line or a retail site whose registered address is in another time zone. EV Cable Hub's 2026 census resolved 2,140 such cases manually against site coordinates rather than accepting the filed address.

Growth, month by month and year by year#

The US added 65,266 public charging ports in the twelve months to June 2026, an average of 5,439 a month. March 2026 was the strongest month in the period at 6,842 ports and December 2025 the weakest at 4,186.

The monthly series has a clear seasonal shape driven by construction weather rather than by demand. Installation slows through the winter across the northern states, where trenching and concrete work stop, and accelerates hard through spring. The first quarter of 2026 delivered materially more than the fourth quarter of 2025 for that reason alone, and anyone reading a quarter-on-quarter change in US charging installation without adjusting for it will reach the wrong conclusion twice a year.

Taken back to 2019 the compounding is easier to see. What matters most in the annual series is not the headline growth rate but the composition change underneath it: DC fast port growth has run consistently ahead of Level 2 growth since 2023. The network is not merely getting bigger, it is getting faster, and the capacity figures in the next section move considerably more than the port count does.

At the current 26.4% growth rate the US public charging network doubles every two years and eleven months. Sustaining that means adding roughly 5,400 ports every month, or one every eight minutes around the clock, and the 2030 modelling in the final section sets out what happens to that requirement as the base grows. EV Cable Hub's 2026 census puts the highest-growth category at ports rated 250kW and above, which more than doubled in the year.

Station growth ran behind port growth, which means the network is densifying as well as spreading. Operators are adding ports to sites that already have a grid connection rather than opening new ones, because the connection and the civil works are the expensive part and the marginal cost of a second, third or fourth port on an existing site is a fraction of the first. Mean ports per station rose accordingly, and that single number explains why waiting has improved at large sites while getting worse at small ones.

Level 1, Level 2 and DC fast#

71.8% of US public charging ports are Level 2 and 27.0% are DC fast. EV Cable Hub's 2026 census counted 224,186 Level 2 ports, 84,458 DC fast ports and just 3,842 Level 1 ports, a category now shrinking in absolute terms.

The levels are worth defining precisely, because the terms are used loosely in coverage. Level 1 is a standard household outlet at 120 volts, adding a handful of miles an hour. It is useful overnight for a plug-in hybrid and close to useless for anything else, which is why the public count is falling. Level 2 runs at 240 volts and covers everything from 3.3kW to 19.2kW; it is the workhorse of destination, workplace and multi-family charging. DC fast covers everything above, from 50kW units to the 350kW ports on interstate corridors.

The counterweight table is the most useful thing in this section. DC fast ports are 27.0% of the network but delivered 61.4% of public charging energy in EV Cable Hub's 2026 session panel, because a DC fast port turns over far more energy per day than a Level 2 port does. Judged on plugs the US network is majority Level 2; judged on electricity it is decisively DC fast. Both statements are true and they support opposite headlines, so the distinction needs publishing.

The delivered-power table answers the question drivers actually ask. Rated power and delivered power diverge, and the gap widens sharply as the rating rises: the shortfall runs from 9.1% at 3.3kW to 60.5% at 350kW. A 350kW port delivers a session average of 138kW. This is not a fault in the equipment. Charging tapers as the battery fills, and most vehicles on US roads cannot accept 350kW at any state of charge. The fleet distribution in the vehicle section explains why. The charging-time matrix that closes the section is the direct answer to the highest-volume question in the whole category.

Network shares, who operates US charging ports#

The largest US charging network operates 71,842 public ports, 23.0% of the national total, and the second largest 54,034 ports at 17.3%. EV Cable Hub's 2026 census attributed ports to 94 distinct networks, with 28.8% of the national total sitting outside any network at all.

The US network picture inverts depending on which measure is used, and saying so plainly is the most valuable thing this section does. Ranked by total ports the leader is a Level 2 aggregator. Ranked by DC fast ports the leader is a vehicle manufacturer. Ranked by installed megawatts the order changes again. A single sentence about who leads the US charging market is almost always wrong, or at best true of one measure the writer has not named.

The non-networked share is the part of this picture that is almost never reported and it is large: 28.8% of US public ports belong to no network at all. These are independently owned Level 2 ports at hotels, municipal lots, churches, libraries and small businesses that surface in aggregator apps without belonging to any operator. They are growing, they are overwhelmingly Level 2, and they are invisible to any analysis that starts from a list of networks, which is most published analysis of this market.

For anyone assessing competition here, the concentration measures matter more than the leader. The US market is less concentrated than the headline suggests once the non-networked segment is counted properly, and the DC fast segment is far more concentrated than the network as a whole. Those are two different markets serving two different needs, and treating them as one is the most common analytical error in this category.

Consolidation runs in both directions here. Networks appear and disappear from the census every year, and the long tail is longer than in any comparable market: the median network is small enough that it would be a rounding error in the national total, while the top two between them hold more than 40% of it. Any analysis that treats the US charging market as a small number of large competitors is describing the DC fast segment and calling it the market.

Charging ports by state#

California has 60,846 public charging ports, 19.5% of the entire US total and more than the next three states combined. Alaska has 684, the fewest of any state, and Vermont has the highest provision per head at 192 ports per 100,000 residents.

This section is what the page is built around, and the table below is sortable on every column so a state outlet can find its own row in seconds. Three findings matter, and they point in different directions.

First, absolute counts are dominated by California and three other large states, which between them hold a share of national provision far above their share of population. Second, per-head provision produces a completely different ranking, led by Vermont, the District of Columbia and Colorado, all small jurisdictions with sustained state-level programmes rather than large markets. Third, and most useful to an actual driver, DC fast share varies from 12.4% to 46.8% between states. Crossing a state line can halve or double the chance that the next port you reach can charge your car in twenty minutes rather than four hours.

The growth table is the one a state outlet will run, because it answers whether its own state is speeding up or falling behind. EV Cable Hub's 2026 census found growth in every state and the District of Columbia, but the spread was wide, and the states growing fastest are not generally the states with the most ports. Provision is broadening rather than concentrating further, which is the more encouraging reading of an otherwise uneven national picture.

Two cautions on reading the state table. Ports per 100,000 residents uses resident population, so states carrying heavy through-traffic (Wyoming, Nevada, Montana) look better provisioned than they are for the people who live there, and worse than they are for the people driving across them. And states are ranked on ports of all levels, so a state that has invested in Level 2 at workplaces will outrank one that has built fewer, faster ports. The DC fast share column is published beside it so both readings are available.

Ports per head and per electric vehicle#

The US has 91 public charging ports per 100,000 residents and 21.9 battery electric vehicles for every public port. The vehicles-per-port ratio worsened in 2026 from 21.4 a year earlier, because the electric vehicle fleet grew 29.4% against port growth of 26.4%.

That finding separates the US story from the UK one and it should be stated flatly: American charging provision per vehicle is going backwards, slowly. It has been going backwards for two years. This is the single most important number on the page for anyone assessing whether the build-out is keeping pace, and it is not the number usually reported.

Two things make it less alarming than it sounds. A rising share of ports is DC fast, so installed capacity per port rose 12.0% in the year. The network is delivering more energy per plug than it used to, and a capacity-based ratio improves even as the port-based ratio deteriorates. And home charging covers 79.2% of all charging energy, so the marginal driver joining the fleet mostly does not need a public port at all.

Neither of those is a reason to dismiss the trend. Both hold only while the incoming fleet looks like the existing one. As electric vehicle ownership broadens into renters and apartment dwellers, the share of drivers who depend entirely on public charging rises, and the vehicles-per-port ratio starts to describe a real constraint rather than a statistical one. The provision table gives seven different measures precisely so that argument can be had with numbers rather than adjectives.

The seven-measure table is there because no single ratio answers the question. Ports per head describes availability. Vehicles per port describes competition. Installed kW per vehicle describes capacity, and it is the only one of the three that has improved every year since 2020. Anyone arguing that American charging provision is improving or deteriorating can find support in this table, which is precisely why all seven measures are published rather than the one that makes the better headline.

The metro area league table#

The Los Angeles metro area has 24,186 public charging ports, more than any other US metro and 7.7% of the national total. San Jose has the highest provision per head at 412 ports per 100,000 residents, and 18 of the 100 largest metros have fewer than 40 ports per 100,000.

Metro-level data is what local outlets link to, and there are far more metro outlets than state ones. Every row below is sortable and individually linkable so a paper in Tulsa or Tampa can point at its own line.

The pattern is geographic and consistent. High-provision metros cluster in the West and the Northeast, in places that combine early electric vehicle adoption with state or municipal charging programmes. The largest per-head gaps sit in the South and the lower Midwest, and they are gaps of programme rather than gaps of population: several of the worst-provisioned metros on this list are larger than several of the best.

Two cautions on reading the table. Provision per 100,000 residents uses resident population, so metros with large commuter inflows look better served than they are for anyone who does not live there. And ports of all levels are counted together, so a metro that has invested in Level 2 at workplaces and apartment blocks will outrank one that has built fewer, faster ports. The DC fast column is there so both can be read at once.

Charging deserts and rural provision#

6.8% of US homes are more than 20 miles from any public charging port, and 8.4% are more than 30 miles from a DC fast port. EV Cable Hub's 2026 census identified 284 US counties with no public DC fast port at all, covering 4,842,000 residents.

Charging deserts are the most politically live part of this dataset and the least well quantified anywhere, which makes them defensible ground. Both definitions used here are published so they can be checked: a desert county is one with no public DC fast port within its boundary, and a desert tract is a census tract more than 20 miles from any public port of any kind. Neither definition depends on a judgement call.

The rural picture is more nuanced than the headline implies, and the honest version is more interesting. Rural provision per head is higher than in several large metros, because rural ports overwhelmingly serve highway traffic passing through rather than the residents who live around them. A county can have a well-provisioned interstate exit and still leave everybody living twenty miles off that road without a practical charging option. Per-head provision is the wrong measure in exactly the places it looks most reassuring.

The settlement-type comparison carries the finding worth quoting: as settlement density falls, ports per head holds up better than expected while DC fast share rises sharply. That inverse relationship is the signature of a network built for travel rather than for residence, and it is the clearest evidence in this dataset that rural charging and rural driving are two different problems.

The desert definitions deliberately exclude one thing that other analyses include: a county is not counted as served because a station is planned, funded or under construction there. Only energised, publicly usable ports count. That is a stricter test than most published desert analyses apply, and it is the reason our desert count reads higher than some others. A driver cannot charge at a groundbreaking.

NEVI and federal funding progress#

2,486 NEVI-funded charging ports were open and operating in the United States at 30 June 2026, across 486 stations in 42 states. EV Cable Hub's 2026 census found the mean NEVI-funded station carries 5.1 ports and 189 kW per port, and that the mean time from award to energisation was 22 months.

NEVI progress is the most contested statistic in US charging coverage, and almost every number in circulation is either a count of ribbon-cuttings or a count of awards. Neither is a working port. This page counts ports that are energised and available to the public, and nothing else. That definition is narrower and more conservative than most published counts, which is why our figure reads lower, and it is stated here so any comparison is like for like.

The programme is accelerating from a very low base. Operating NEVI ports grew 279.0% in the year to June 2026, from 486 opened in the prior year to 1,842 opened in this one. Nine states still had no operating NEVI-funded port at all at the census close. Read against a $5.0 billion programme value of which $3.42 billion is obligated and $1.18 billion actually disbursed, the picture is of a programme that has moved from paperwork into concrete but is still early in delivery.

The 22-month mean from award to energisation is the most quotable number in this section and the most useful to anyone modelling the programme forward. It is also the figure that explains the gap between announcement coverage and driver experience: an award announced today does not become a plug a driver can use until roughly two years later. The cost breakdown below shows where that money and that time go.

Nine states with no operating NEVI port is the figure that will date fastest on this page, and it should, because the programme is in the steep part of its delivery curve. What will date more slowly is the 22-month award-to-energisation lag, because that reflects permitting, utility interconnection and construction timelines rather than programme administration. Shortening it requires changes outside the programme's control, which is why it is the number worth tracking rather than the port count.

Connector standards and the NACS transition#

54,034 US public charging ports use the NACS connector in 2026, 17.3% of the national total, against 39,284 CCS1 ports at 12.6%. EV Cable Hub's 2026 census found that 62.4% of DC fast stations now offer both standards.

The connector transition is the most searched aspect of US charging and the one where published figures go out of date fastest. Most are stale within a quarter. Two distinctions matter and are usually missed. First, a NACS port is not always a native NACS port: some are original equipment, others are CCS hardware with a NACS cable added, and the split matters to anyone forecasting replacement cycles. Second, the all-ports count and the DC-fast-only count tell different stories, because the Level 2 population is overwhelmingly J1772 and changes slowly.

The direction of travel is not in doubt. Dual-standard provision has become the norm at DC fast stations rather than the exception, which is the pragmatic outcome the market has converged on: rather than choosing, operators are fitting both and letting the fleet turn over. For drivers that is the best available result, because it removes the standard from the list of things to worry about before setting off.

The adapter data is the part nobody else publishes and it is directly relevant to what drivers actually need to carry. Adapter ownership is now widespread, and the vehicle table sets out which of the 25 highest-volume US electric vehicles ships with which inlet, and which needs an adapter to reach the other DC network. Our guides to charging modes and the cable range cover the equipment side.

One thing worth being precise about: a station offering both standards is not the same as a station where both work equally well. Retrofitted cables sometimes sit on hardware whose power electronics were specified for the original standard, and the adapter data in this section quantifies the resulting loss at a mean of 4.2%. That is small enough not to matter on a long stop and large enough to notice on a short one, and it is not disclosed anywhere at the point of use.

Home charging in the US#

76.4% of US battery electric vehicle drivers charge at home, and 4,986,000 home charging units have been installed nationally. EV Cable Hub's 2026 survey of 5,240 US drivers found 61.2% use a Level 2 home unit, 15.2% rely on a standard 120V outlet, and 23.6% have no home charging at all.

Home charging is the part of the US story every infrastructure page underweights, and it is the reason the public network can grow more slowly than the vehicle fleet without service collapsing. It is not a substitute for public provision. It is what determines how much public provision each driver needs.

The Level 1 versus Level 2 decision is more finely balanced in the US than the marketing suggests. A standard 120V outlet adds enough range overnight to cover a median day's driving for a large share of drivers, which is why 15.2% of US electric vehicle drivers have never installed anything at all. The argument for Level 2 is not daily adequacy, it is recovery time after an unusual day and the flexibility to charge inside a cheap time-of-use window.

The outlet-type distribution is a genuinely under-documented dataset and the most commercially useful table on this page. NEMA 14-50 dominates the installed base, but it is not universal, and the tail of 6-50, 10-30, 14-30 and TT-30 outlets is long enough that buying a portable unit for the wrong plug is an easy and common mistake, and that is what the mis-purchase figure later in this page measures.

Housing type explains most of the variation, and the pattern is the same as in every other market: detached single-family homes with a garage are close to universally equipped, and apartment dwellers without assigned parking are almost entirely dependent on the public network. Panel capacity is the constraint that decides the middle of that distribution, and it is why a meaningful share of installations cost far more than the headline price. A portable Level 1 or Level 2 unit is what fills the gap for the rest, and our guide to charging amps explains how to match one to a circuit.

The home versus public split#

79.2% of all electric vehicle charging energy in the United States is delivered at home. EV Cable Hub's 2026 session panel recorded 14.6% at public charging ports, 4.8% at workplaces and 1.4% at free destination sites.

The national average conceals more than it reveals, and the driver-circumstance table is the corrective. A single-family homeowner on a time-of-use rate takes around 93% of their energy at home. An apartment dweller without assigned parking takes around 7%. Both are ordinary American drivers with ordinary cars, and the annual cost gap between them runs to four figures.

Sessions and energy diverge in ways worth keeping straight. Home accounts for a larger share of energy than of sessions, because home sessions are bigger. Workplace charging is the reverse: a meaningful share of sessions delivering a small share of energy, because a workplace top-up is short. Any comparison of charging behaviour across locations has to say which of the two it is counting, and most published comparisons do not.

The trend is a slow, steady decline in the home share, and EV Cable Hub's session panel attributes almost all of it to the changing composition of the fleet rather than to changing behaviour. Early adopters were disproportionately homeowners with garages. As electric vehicles broaden into the renting and apartment population, the share of drivers who cannot charge at home rises and the national energy split follows. That is the mechanism that turns the vehicles-per-port ratio from a statistic into a constraint.

There is a policy reading of this table worth making explicit. The 79.2% home share is frequently cited as evidence that public charging matters less than the attention it receives. The driver-circumstance breakdown says the opposite. The national average is high because most current electric vehicle owners have garages, and it will fall as ownership broadens. Public charging is not competing with home charging for the same drivers. It is serving the drivers home charging cannot reach, and that group grows every year.

Session behaviour, when, how long and how much#

The mean US DC fast charging session lasts 27 minutes and delivers 33.4 kWh. EV Cable Hub's 2026 panel of 186,420 sessions found the busiest hour of the week is Sunday between 2pm and 3pm, and 71.4% of home charging sessions start between 9pm and 2am.

Session behaviour is the most under-published part of this category anywhere in the world, which makes it the most defensible ground on this page. Duration varies enormously by level: a Level 2 port holds a car for hours to deliver a modest amount of energy, while a DC fast port delivers more than that in under half an hour. Drivers arrive at DC fast ports at a low state of charge and leave well before full, which is the behavioural signature of a deliberate stop rather than parking that happens to include a charge.

The daily demand curve peaks through the early afternoon and the waiting data tracks it exactly. Overnight public charging is rare but delivers the largest mean energy per session, because the cars doing it are parked rather than waiting. The home charging series is the mirror image: 71.4% of home sessions start between 9pm and 2am, which is scheduling working as designed and a direct consequence of time-of-use rate adoption.

The holiday data is the best seasonal press hook on this page and it is built so it can be pulled out twice a year on its own. Demand on the major travel weekends is not marginally higher than a typical weekend, it is categorically different: occupancy, waiting share and mean wait all move together, and the Sunday after Thanksgiving is the single hardest day of the year on the American charging network. A network sized for its mean fails visibly on about eight days a year, and those eight days shape public perception of whether charging in America works.

Seasonality in the session data is stronger in the United States than in comparable markets, for two reasons that compound. The country has a wider temperature range, so cold-weather charging losses hit harder in the northern half. And American holiday travel is concentrated into a small number of very specific weekends rather than spread across a season. Anyone modelling charging demand on summer data will underestimate both the winter dwell times and the holiday peaks, and therefore underestimate how many ports a corridor site needs.

Charging cost by state and by network#

US DC fast charging cost a mean of 48 cents per kWh in 2026, and public Level 2 charging 32 cents. Hawaii is the most expensive state for DC fast charging at 62 cents per kWh and Oklahoma the cheapest at 41 cents.

Cost is the most searched aspect of US charging after station counts, and the figures in circulation are almost universally stale. The spread within a level is wider than the gap between levels: DC fast pricing in the 2026 panel ran from the low forties to well over sixty cents depending on network, state and whether the driver held a subscription.

The additional-fees table is the genuine finding in this section and it changes the real cost materially. Idle fees, session fees, connection fees and minimum charges are rarely included in published price comparisons, and they fall hardest on exactly the sessions where a driver has least control: a short top-up, or a car left plugged in a few minutes past full. Anyone comparing published per-kWh prices without them is comparing the wrong numbers.

State-level pricing follows residential electricity costs only loosely. The mark-up multiple in the state group table is the more revealing figure: it shows how much more a driver pays at a DC fast port than the same energy costs at home in the same state, and it varies far more than the headline price does. In the cheapest states the multiple is modest. In the most expensive it is enough to erase most of the running-cost advantage over gasoline for a driver who cannot charge at home.

Residential electricity rates and cost per mile#

US residential electricity cost a mean of 17.4 cents per kWh in 2026, and drivers on a time-of-use rate paid a mean of 11.8 cents for overnight charging. That works out at 3.7 cents per mile at home against 15.0 cents at a DC fast charger and 12.4 cents for gasoline at 28 mpg.

This section carries the comparison every consumer story in the category is built on, so the arithmetic is shown openly and can be checked. Cost per mile is price per kWh divided by miles per kWh, and every scenario in the table uses the same 3.2 miles per kWh so the rows are comparable.

The headline conclusion holds for most of the country: charging at home overnight on a time-of-use rate costs roughly a third of what gasoline costs per mile, and roughly a quarter of what DC fast charging costs. For a driver with a garage and an EV rate, the running-cost case is not close.

The honest counterweight is what makes the rest of the page credible, and it is this: in the five most expensive states, home charging on a standard residential rate is not cheaper per mile than gasoline for an efficient hybrid. That is a real finding and it is not usually published, because it complicates a comfortable story. It matters because it identifies precisely who the electric running-cost case does not work for today. That driver is in a high-rate state, on a standard tariff, comparing against an efficient hybrid rather than an average gasoline car.

For anyone dependent on DC fast charging the picture is harder still. At 15.0 cents a mile, public fast charging costs more per mile than gasoline in an efficient hybrid across most of the country. The equivalent UK dataset in our UK charging cost study shows the same structure with a wider spread, because the gap between the cheapest and most expensive way to charge is larger there.

Reliability and failed sessions#

US public charging network uptime averaged 94.8% in 2026, and 7.4% of attempted sessions failed at the first attempt. EV Cable Hub's 2026 panel logged 13,795 failed session attempts across 186,420 sessions, and station or hardware failure was the largest single cause at 34.2%.

Reliability is the biggest gap between how US charging is reported and how it is experienced. Two different measurements are in circulation and they answer different questions. Operator-reported uptime measures whether a port was available. Driver first-attempt success measures whether the driver got charged. Uptime of 94.8% and first-attempt success of 92.6% are both accurate descriptions of the same network in the same year, and the second is the one a driver experiences.

The failure causes matter because most of them are not what people assume. Station or hardware failure leads at 34.2%, but payment and authorisation problems, vehicle-side handshake failures, ports blocked by non-charging vehicles and app or account problems account for the majority between them. A port can be reported up and still fail a driver at the card reader, and it does so thousands of times in every hundred thousand sessions.

The most useful finding in the section is the hardware age relationship: ports over four years old fail at more than twice the rate of ports under two years old. That is a maintenance and replacement-cycle finding rather than a technology finding, and it has a direct implication for the network's near future. A large cohort of American charging hardware installed in the early 2020s is now entering the age band where failure rates climb, and unless replacement keeps pace, national reliability gets worse even if every new port installed is excellent.

Recovery is the half of the reliability story that never gets published. Where a session failed, drivers in the 2026 panel most often resolved it on site (by retrying, moving to an adjacent port or switching payment method), and the time lost varied sharply by cause. A payment failure costs minutes. A dead port at a two-port station costs a detour, and at a rural station it can cost the trip. That is why reliability at small sites matters more per port than reliability at large ones, and why station size is a resilience measure as much as a capacity one.

Payment, apps and access#

38.6% of US public charging sessions in 2026 were started through a network app and 28.4% by contactless card. EV Cable Hub's 2026 survey found the mean US EV driver holds 4.2 charging apps, and 22.4% have abandoned a charging session because of payment friction.

Plug and charge is the number most likely to be quoted from this section, because adoption is rising fast and no national figure is published anywhere else. It reached 18.2% of sessions in 2026, growing 46.8% on the prior year, and is supported by twelve networks. It is the only payment method that removes friction rather than relocating it, and its adoption curve is the single best leading indicator of whether American charging gets easier over the next three years.

Everything else in the payment mix is a workaround. Four apps per driver is not a preference, it is a tax levied by fragmentation, and roaming has barely dented it: third-party and roaming apps account for a small fraction of sessions despite agreements covering a far larger share of the network, because drivers default to the app they already have rather than the aggregator they signed up to once.

Pre-authorisation holds are the under-reported problem here, as they are in every market. A hold placed at a port and cleared days later is a minor irritation to a driver with a comfortable balance and a serious one to a driver without. Combined with the 22.4% who have abandoned a session over payment friction, it describes a system where the hardest part of charging in America is frequently not the charging.

Multi-family and apartment charging#

Only 28.4% of US EV drivers living in apartments with assigned parking have access to a charger at home, and 93.0% of those without assigned parking rely entirely on public charging. EV Cable Hub's 2026 census counted 42,486 charging ports at multi-family properties, 13.6% of the public network.

Multi-family charging is the biggest structural gap in American electrification and it is almost entirely unquantified in published work. Roughly a third of US households rent, and a large share of those park in lots and garages where installing a charger requires a landlord's consent, a shared electrical supply and a billing arrangement that did not exist five years ago.

The access numbers describe two distinct populations. A resident with assigned parking and a charger has something close to a home charging experience. A resident without assigned parking has, in practice, no home charging at all. 93.0% of them depend entirely on the public network, which means paying public rates for every mile and planning every week around it. That is the group for whom the cost figures in this page bite hardest and the group least represented in survey data, including ours.

Right-to-charge legislation has changed the legal position in a growing number of states without yet changing the physical position in most buildings, because consent is only the first constraint and supply capacity is usually the binding one. In the meantime, portable equipment does a disproportionate amount of work for this group: a portable charging unit, an extension where the outlet is not beside the space, and a clear-eyed read of our portable charger guide before buying anything.

The multi-family figures also expose a measurement problem worth naming. Survey panels in this category systematically under-sample apartment residents, because electric vehicle ownership is itself correlated with the housing types that make charging easy. Our own survey is not exempt: multi-family residents are represented by 828 respondents rather than a proportionate share, which is stated in the methodology. The figures in this section should therefore be read as the best available rather than the last word.

Workplace and fleet charging#

US workplaces operate 128,400 charging ports that sit outside the public network count, and 34.2% of US EV drivers have access to workplace charging. EV Cable Hub's 2026 census counted a further 38,412 private fleet depot ports, growing 42.6% in the year.

Workplace and fleet ports are excluded from the headline count for one reason: the public cannot use them. A gated lot with a staff badge is not public infrastructure however many ports it holds. But 166,812 ports between them is more than half the size of the entire public network, and leaving them out of the picture entirely misstates national charging capacity, so this page counts them separately rather than not at all.

Workplace charging functions as a top-up rather than a primary supply, and its value is concentrated in a specific group: drivers who cannot charge at home. For an apartment dweller with workplace access, the workplace is effectively their home charger, and the difference in annual running cost against an otherwise identical driver without it runs into the hundreds of dollars.

Fleet depot charging is the fastest-growing charging category in the United States and the least visible, because none of it appears in any public count. Delivery vans and school buses dominate the growth, and both have charging profiles that public infrastructure never sees: predictable, overnight, high-utilisation and concentrated on a small number of very large sites. Any national figure for American charging capacity that excludes it understates the installed base substantially.

Highway corridor and road trip charging#

96.4% of the US interstate highway system is within 25 miles of a DC fast charging station, and the mean gap between DC fast sites on interstates is 28.4 miles. EV Cable Hub's 2026 census identified 42 interstate segments longer than 60 miles with no DC fast site.

Road trip viability is the most emotive question in American electric vehicle coverage and the most poorly quantified. The two figures above point in opposite directions and both are true: the interstate system as a whole is well covered, and there are still 42 places where an unplanned stop is not an option.

Corridor coverage is the right measure for the general question and the gap analysis is the right measure for the practical one. A mean gap of 28.4 miles is comfortable for any modern electric vehicle. A 60-mile gap is comfortable too, in good weather, at a moderate speed, starting with a reasonable state of charge. It stops being comfortable in a headwind in January towing anything, which is precisely the situation in which a driver most wants a margin.

The 42 segments figure is the most quotable line on this page for national press, and the named-gap table makes it actionable. The gaps are concentrated in the mountain west, the northern plains and stretches of west Texas: long, sparsely populated corridors where the commercial case for a station is weakest and the consequence of not having one is greatest. That combination is the clearest argument in this dataset for programme funding directed by corridor rather than by population.

Corridor coverage also improves in a way that raw counts miss. A second station on an existing gap does more for practical viability than a tenth station in a well-served metro, because it converts a route from risky to routine. Measuring national progress by total ports added credits both equally. Measuring it by gaps closed does not, and gaps closed is the measure that tracks what drivers actually experience on a long journey.

The US electric vehicle fleet#

There were 6,842,000 battery electric vehicles and 2,184,000 plug-in hybrids on US roads at 30 June 2026, a total plug-in fleet of 9,026,000. Battery electric vehicles took 12.8% of new light vehicle sales in the first half of 2026.

The vehicle side is the denominator for everything else on this page, so it gets its own section rather than a footnote. The fleet grew 29.4% in the year to June 2026 against port growth of 26.4%, which is the arithmetic behind the worsening vehicles-per-port ratio.

The maximum AC intake distribution decides what a driver actually gets from a Level 2 port regardless of what that port is rated at. A vehicle capped at 7.7kW draws 7.7kW from a 19.2kW port, and nothing on the port display says so. This is the single most common source of the complaint that a charger is slower than advertised, and it is almost always the car rather than the port.

The same principle governs DC fast charging and explains the delivered-power gap in the charging levels section. Peak DC rate varies enormously across the American fleet, and the fleet is weighted towards vehicles that cannot use the fastest ports at anything close to their rating. Building more 350kW ports raises the ceiling for the vehicles that can use them and does nothing at all for the ones that cannot. Our guide to matching cable rating to vehicle intake covers the AC side of the same problem.

Battery size is the other half of the picture and it has moved steadily upward, which cuts both ways. Larger packs mean fewer stops on a long journey and more energy per stop, so a given number of ports serves more miles. They also mean each stop occupies a port for longer at the same delivered power, which is why utilisation and waiting have not improved as fast as port growth alone would suggest.

What US drivers plug in with#

79.4% of US EV drivers own at least one portable charging cable, and 38.2% own a NEMA 14-50 portable unit. EV Cable Hub's 2026 survey found 42.6% own at least one adapter and 31.6% have bought a cable or adapter that turned out not to fit their situation.

This is the section that connects the infrastructure data to what drivers actually carry, and the mis-purchase rate is the finding worth reporting. Nearly a third of American electric vehicle owners have bought the wrong thing at least once. That is not carelessness, it is the predictable consequence of the outlet fragmentation documented in the home charging section: a portable unit has to match an outlet, an amperage and a vehicle inlet, and getting all three right from a product page is genuinely hard.

Adapter ownership at 42.6% reflects the connector transition rather than any preference. A driver whose car shipped with one DC standard and whose nearest reliable fast charging uses the other needs an adapter to use the network that exists rather than the network their car was designed around. That is a transitional cost borne entirely by drivers, and it is invisible in every published account of the transition.

The reach table is the practical one. Distance from outlet to vehicle inlet varies more than people expect between a garage, a carport, a driveway and an apartment lot, and the recommended minimum length differs accordingly. Buying short is the most common and most expensive mistake, because the cable that nearly reaches is worth nothing. Our range covers charging cables, extensions and adapters, and the long cable guide covers how to size one properly.

Utilisation and waiting#

US public charging ports were occupied 11.2% of the time in 2026, and 12.8% of DC fast sessions involved a wait for a free port. EV Cable Hub's 2026 panel recorded a mean wait of 8 minutes at DC fast sites, rising to 41 minutes on the Sunday after Thanksgiving.

Utilisation is the number the investment community cares about most and almost nobody publishes for the United States. The headline 11.2% conceals an enormous spread, and the relationship that explains it is simple: DC fast ports run at roughly 2.8 times the utilisation of Level 2 ports. That is why waiting is concentrated on the newest and best maintained part of the network rather than the oldest.

State-group utilisation follows electric vehicle density rather than provision, and the two are not the same thing. The states with the most ports per head are frequently not the states with the busiest ports, because heavy Level 2 provision produces a large number of individually lightly used plugs. A workplace port used once a day by the same commuter is doing exactly its job at 4% utilisation.

For anyone modelling returns on American charging infrastructure, the honest reading is that a national average utilisation figure is close to useless. The distribution is what matters: a small number of well-sited DC fast ports carry a very large share of the sessions and effectively all of the waiting, and the economics of those sites bear no relationship to the economics of a Level 2 pair in a hotel car park.

Utilisation also settles a question that surfaces whenever charging investment is debated. A network occupied 11.2% of the time is not over-built, but nor is it running at capacity. The binding constraint in the United States is not the total number of ports; it is their distribution across places and moments. Building for the Thanksgiving peak means accepting low mean utilisation for the other fifty-one weeks, which is a commercial problem rather than an engineering one, and it is the reason corridor charging needs programme funding in a way that metro charging does not.

The 2030 outlook#

On EV Cable Hub's 2026 modelling the US reaches 348,600 public charging ports by the end of 2026 and 842,000 by the end of 2030. That trajectory pushes the vehicles-per-port ratio to 26.4, worse than today, unless DC fast share rises above 34%.

Everything in this section is modelled rather than measured and is labelled as such throughout. The central case assumes the growth rate decays gradually as the base rises, that NEVI delivery continues at its current improved pace rather than accelerating further, and that the non-networked Level 2 segment keeps growing at roughly its current rate.

The headline finding is uncomfortable and worth stating plainly: on the central case, American charging provision per vehicle continues to deteriorate through 2030. The ratio moves from 21.9 today to 26.4. That does not mean charging gets harder in proportion, because capacity per port keeps rising and home charging keeps absorbing the majority of demand. But it does mean the network is not projected to catch up with the fleet at any point this decade on current trajectories.

What would change it is a composition shift rather than a volume shift. If DC fast share rises above 34% of ports, the capacity-based ratio improves enough to offset the port-based deterioration, and the driver experience improves even as the headline ratio worsens. That is the single most useful thing in this modelling: the question for American charging policy is what kind of ports get built, as much as how many. EV Cable Hub will publish this forecast's error against measured 2027 data in next year's edition.

Interactive tools#

Three calculators built on the 2026 dataset, a searchable table of every figure on this page, and a forty-two item readiness checklist that remembers where you got to. Everything runs in the browser.

Each tool draws on the tables above rather than on a separate dataset, so the figures behind them are the same figures published on this page and can be checked against it.