What to Do When a Public EV Charger Won't Start

One in seven public charging visits in the US in 2025 ended with the driver getting nothing for their trouble — no kilowatt-hours, sometimes a pending charge on their card anyway, and a dead stall between them and the next 100 miles [1].

By Liam Whitcombe, EV Ownership & Running-Cost Analyst · Published August 11, 2026 · Data current to Q3 2026

---

A charger that won't start is not a rare edge case. J.D. Power's 2025 survey of 7,428 owners found 14% of public charging visits ended without a successful charge, and 60% of those failures were the charger being out of service rather than anything the driver did wrong [1]. ChargerHelp's independent dataset of over 100,000 sessions found first-time success at just 71%, even at stations whose operators self-report 98%+ uptime — a gap that exists because "uptime" measures whether a charger is technically online, not whether a specific driver's plug-in actually worked [2][3]. This guide is built around that gap: what to try at the plug, when to stop trying, how to get the fault on record with the network that owns the charger, and what each of the five major US networks actually says — not implies — about refunding money for electricity you never received.

Try this five-step sequence before you write the charger off

Unplug and replug is the correct first move, and it works more often than it should. Electrify America's own support guidance opens with exactly this: confirm the connector clicked fully into the port, and if it didn't, "remove the cable and plug in again" [9]. A loose or partially seated connector is one of the most common reasons a session fails to initiate, and it costs nothing to rule out.

Second, restart the session through the network's app rather than tapping your card or fob again. Card and RFID taps often just resend the same authorization request that already failed once, while an app-based stop-and-restart clears the session state on the network's backend and re-establishes the handshake with the charger from scratch. This is consistent with why the Joint Office's ChargeX Consortium — the DOE-backed group of national labs and network operators formed to fix exactly this kind of friction — has been pushing networks toward automated retry logic so a driver doesn't have to manually cycle a session at all [25].

Third, if the connector or stall itself seems to be the problem, try a different plug at the same station before you leave. Multi-stall sites are common precisely because a single connector, cable or power module can fail independently of its neighbors — the UC Berkeley field study of Bay Area chargers found failures clustered in specific non-functioning connectors (blank screens, payment failures, initiation errors) rather than whole-station outages, with cable-length problems a smaller but real factor at 4.9% of connectors [4]. A dead port two feet away often sits next to a working one.

Fourth, look for an error code on the charger's own screen before you give up. This detail matters more than it used to: the ChargeX Consortium's national labs released a formal set of 26 minimum required error codes specifically because charger manufacturers, car makers and network operators were reporting the same underlying fault in different, incompatible language, which slowed diagnosis on both sides of a support call [28]. A code you can read off to a support agent — even if you don't know what it means — gets you past the "have you tried unplugging it" stage of the call faster.

Fifth, know when not to force it. If a connector, cable or holster shows visible damage — cracked housing, exposed wiring, scorch marks, a connector that won't latch because something inside is broken — do not keep trying to plug in. That is a stop point, not a retry point: report it and move to another charger. None of the troubleshooting above should be attempted on hardware that looks physically compromised.

Consumer Reports' own survey of 1,230 EV and plug-in hybrid owners found that 19% of charging problems involved the charger not starting even though payment had already been accepted — a distinct and worse failure mode than simply not starting, because now there's a pending or posted charge with nothing to show for it [34]. If that happens to you, the five-step sequence above still applies first, but keep a screenshot of the payment confirmation or the error screen the moment it appears; it is the single most useful thing you can hand a support agent, and, as covered further down, often the deciding factor in whether a refund conversation goes anywhere.

When to stop retrying and drive to another charger

Two to three attempts, five to ten minutes total, is a sensible ceiling. Past that point you are very likely dealing with a hardware fault rather than a transient software glitch, and J.D. Power's data backs that instinct: 60% of failed visits in 2025 were attributed to the charger simply being out of service, not a fixable in-the-moment error [1]. Continuing to retry a genuinely broken unit doesn't just waste your time — on a busy corridor site it also occupies a stall other drivers need, the exact behavior the idle and congestion fee system exists to discourage, even though a non-charging car isn't actually accruing those fees.

The decision to abandon a stall and drive on is where PlugShare earns its place in the glovebox, and it's worth checking before your next stop rather than only after this one fails — more on exactly how that works below. For now, the practical rule is: if the five-step sequence above doesn't produce a charging session within about ten minutes, report the fault (next section) and route to the next-nearest station rather than camping on a dead unit. ChargeCostLab's own modeling, detailed later in this piece, puts a typical detour at roughly 8 miles and $0.70 a mile in vehicle cost plus time — real but small money next to the alternative of sitting at a charger that was never going to work.

How to report a broken charger, network by network

Reporting a broken charger takes under 60 seconds on every major US network, but the exact steps differ by operator: Tesla and ChargePoint route through their driver app, Electrify America and EVgo through a 24/7 phone line, and Blink through an online form. The table below is each network's own stated channel, current as of Q3 2026.

Network	How to report	Channel details
Tesla Supercharger	In-app report feature	Open the Tesla app (v4.39.1+), select the location, choose a category, attach a photo; support reviews reports of onsite issues including access, cleanliness and damage [16]
Electrify America	Call Contact Center, or app	Call the number posted above the charger screen, or in-app tap Support > Report Issue; Contact Center is staffed 24/7 at 1-833-632-2778 [9]
EVgo	In-app chat, or phone	Chatbot with photo and the charger's name/asset ID; Call Center answers 24/7/365 at 1-877-494-3833 [10][11]
ChargePoint	Driver app	Chargers > select port > scroll to Report a Problem > choose issue > add comment > Submit; phone support 24/7 at 1-888-758-4389 [13]
Blink	Online report form, or phone	Online "Contact Customer Support" form on the Driver Support page (name, email, issue category, station serial number), or the driver-support line; phone staffed 24/7 at 1-888-998-2546 [19][20]

Sources as marked. Report the fault even if you've already routed to another charger — the network can't dispatch a technician to hardware it doesn't know is down.

A structural point worth understanding: ChargePoint's own FAQ notes that most of its stations are independently owned, so when ChargePoint is notified of a malfunction it "contact[s] the owner and help[s] facilitate the repair" rather than fixing it directly [13]. That's a materially different repair path than Tesla or Electrify America, both of which own and operate their own hardware — it's part of why ChargePoint and EVgo idle-fee and support policies vary by individual site rather than following one company-wide rule, a pattern also documented in ChargeCostLab's comparison of network pricing structures.

Reporting matters beyond your own session. The federal ChargeX Consortium's payment-reliability work explicitly grouped charger failures into five categories — network-related failures, integration and installation problems, hardware robustness, customer-experience friction, and maintenance gaps — and found that consistent driver reporting through exactly these network channels is part of how operators actually locate patterns worth fixing rather than one-off complaints [26]. A report that goes nowhere for you individually still adds to the record a network uses to prioritize truck rolls.

Refunds and compensation: what each network actually owes you

A refund for a failed session is never automatic on any of the five major US networks — Tesla, Electrify America, EVgo, ChargePoint and Blink all resolve it case by case through a support conversation, and the terms of that conversation differ sharply by operator. That case-by-case posture is itself a finding worth sitting with: no operator has chosen to compete on a published, guaranteed make-good for a paid-but-failed session, even as reliability data has become a bigger part of how networks are compared and reviewed.

EVgo is the most explicit, and the least generous on paper. Its terms of service state plainly: "All Charges paid by you are final and non-refundable, unless otherwise determined by EVgo" [12]. That "unless otherwise determined" clause is the whole game — EVgo's help center describes specialized agents working billing and refund requests, typically resolving them within about 48 hours, but the starting legal position is that you are owed nothing and a refund is EVgo's discretionary call [10][11].

Electrify America doesn't commit to a refund policy for failed sessions at all. Its mobile FAQ addresses a related but different question — clarifying that a temporary authorization hold "is not an additional charge" and that drivers are "only charged for the actual cost of your session... after it ends" — but stops short of promising money back if the session never happened [9]. In practice this means the outcome depends on the individual support interaction; drivers who reach a phone agent with a clear description of the fault report mixed results.

ChargePoint and Tesla both resolve billing disputes through support rather than a published rule. ChargePoint's own FAQ answer to "how do I get a refund" is a single line directing you to contact support if you feel you were charged in error, with no stated timeframe or automatic-refund commitment [14][15]. Tesla has no public refund policy for a routine failed session either, but has a track record of correcting billing at scale when a systemic bug is found — it has run mass automatic refunds for incorrectly billed Supercharger sessions in specific regions [17], and was compelled by a class-action settlement to refund 100% of wrongly charged idle fees to affected "free unlimited Supercharging" owners [18]. That history suggests Tesla will make good on a clear, provable billing error, but a single driver's single failed session is still a support-ticket conversation, not an automatic credit.

Blink's record is the most mixed of the five, with driver reports ranging from full refunds to flat refusals depending on the agent and the case; the company directs billing and payment disputes to its driver-support channel and a dedicated payments inbox rather than stating a refund policy on its public site [19][20]. This is also the network category where ChargeCostLab's guide to hidden fees beyond the kWh rate is worth reading before you dispute anything — session fees, activation fees and idle fees can all appear on the same receipt as an energy charge, and knowing which line item you're actually disputing speeds up the conversation.

The pattern across all five is consistent enough to act on: no network guarantees a refund in writing, but every network's support channel can grant one, and the deciding factors are the same regardless of operator — contact support promptly (same day, ideally same hour), have the session ID, the charger's location or asset number, and a timestamped photo of the error screen or receipt ready, and ask specifically for a refund rather than just reporting a fault. A report alone gets a technician dispatched; a report plus an explicit refund request is what actually gets money back.

PlugShare and other apps: check before you drive, not just after

PlugShare's PlugScore is the closest thing available to a pre-drive reliability check, and it works by weighting recent driver check-ins rather than pulling a live technical status feed from the charger itself. When a driver checks in, they mark the visit as a successful charge or flag that they couldn't charge and select a reason — broken charger, payment failure, or the stall being blocked by another vehicle — and each check-in nudges the station's PlugScore up or down accordingly, with more weight on recent activity than on the station's all-time record [21][22]. PlugShare's own guidance is direct about what the score means in practice: a PlugScore of 8–10 indicates recent check-ins were mostly successful, while a score of 5–7 signals mixed results recent enough that you should read the actual comments before committing to that station [21][23].

That's a meaningful distinction to hold onto: PlugScore tells you what other drivers just experienced, not what the network's monitoring system currently reports. The two can and do diverge — it's the same measured-versus-claimed gap driving ChargerHelp's entire argument that self-reported uptime understates real driver friction [2][3]. A station can look "up" on the operator's dashboard and still have a broken screen or a payment terminal nobody's fixed, which is exactly the kind of fault a crowdsourced check-in catches and a uptime percentage doesn't.

Consumer Reports' own guidance points the same direction: check a station's live status in PlugShare or a similar app such as A Better Route Planner before you route to it, not only after a session fails, and keep a backup network account with payment already on file in case the primary charger's card reader or app-link is the thing that's broken rather than the charger itself [34]. None of this eliminates the 14% failure rate — it's a crowdsourced signal, not a guarantee, and a station can fail in the ten minutes between the last check-in and your arrival — but checking PlugScore before a long detour to a specific station is free, takes seconds, and measurably shifts the odds in your favor.

Why chargers fail this often

Reported uptime and actual charge success are measuring two different things, and the gap between them is the whole story. Networks widely report uptime figures at or above 98%, a number that typically means "the charger was online and communicating with the network's backend for X% of the period measured" [2]. ChargerHelp's 2025 analysis of over 100,000 real sessions found first-time charge success at only 71% — meaning nearly three in ten attempts at chargers the network would call "up" still failed for the driver plugging in [2][3]. The same report found success rates fall off with station age: new stations average 85% first-time success, but that figure drops to 69.9% by a station's third year of operation, a decline ChargerHelp attributes to deferred maintenance and aging hardware rather than any change in how uptime is measured [2][3].

Independent field research backs the same conclusion from a different angle. A UC Berkeley study that physically tested 657 Bay Area CCS connectors found only 72.5% fully functional, with 22.7% failing on blank screens, payment errors or initiation faults and another 4.9% simply fitted with cables too short to reach some vehicles' charge ports [4]. ChargerHelp's earlier 2024 dataset, covering more than 19 million data points, found over 26% of stations carried software inaccuracies in their reported status, and that just four networks accounted for 25% of all ports but 75% of the ports that were actually down [5] — reliability problems are not evenly spread across the industry; they cluster.

J.D. Power's driver-facing number folds all of this together: 14% of public charging visits in 2025 ended without a successful charge, an improvement from 19% in 2024 and 20% in 2023, with 60% of those failures caused by an out-of-service charger [1]. Paren's state-by-state reliability index shows the same direction of travel — most states sat in a 90–95% reliability band in Q1 2026, up from a broader 85–92% range a year earlier, though Paren's own reporting stresses that reliability leadership shifts year over year by state and operator rather than settling into a fixed national number, so treat any single "reliability score" as a snapshot, not a guarantee for your specific route [8].

The federal government has set a bar for this: chargers funded under the National Electric Vehicle Infrastructure (NEVI) program must maintain at least 97% uptime per port annually to keep their federal dollars, under the standard published in 23 CFR Part 680 [6][7]. That standard is enforced against the funding recipient — a state DOT or the operator it contracts with — not owed directly to any individual driver, and, as the uptime-versus-success-rate gap above makes clear, meeting a 97% technical uptime bar is not the same thing as a driver actually getting a charge on the first try.

What a failed session actually costs you

Every retry, failed visit and detour has a real cost in time and, often, money — and because none of the major networks publish an industry-wide figure for this, ChargeCostLab built its own model rather than leave the number unquantified.

The assumptions, stated plainly so you can substitute your own: a retry — unplug, replug, restart the session — costs about 7 minutes of friction and happens on roughly 15% of sessions; a fully failed visit, where you leave without charging at all, costs about 30 minutes of wasted time and happens on 14% of visits, matching J.D. Power's measured rate [1]; about half of fully failed visits require an 8-mile detour to the next station at $0.70 per mile in vehicle operating cost; and time is valued at $25 an hour, a mid-range figure for the opportunity cost of an unplanned delay. None of these per-session costs are large individually — that's exactly why they're easy to underestimate.

Run those assumptions across a driver who relies on public DC fast charging for about 120 sessions a year — someone without home charging, or a high-mileage driver — and the retry friction alone costs roughly $79 in time value (120 × 15% × 7 min ÷ 60 × $25), the failed-visit time costs about $70 (120 × 14% × 30 min ÷ 60 × $25), and the detour cost adds about $47 in time plus $55 in vehicle cost (120 × 14% × 50% × [8 mi × $0.70 + (8 mi ÷ 30 mph × 60 min ÷ 60 × $25)]). That totals to roughly $310 a year and 10.5 hours — a full working day and a quarter, lost to chargers that didn't work (our calculation on [1]). A driver who charges mostly at home and only fast-charges on road trips — about 30 sessions a year — scales down proportionally to roughly $78 a year and 2.6 hours (our calculation). Both figures sit well below the annual fuel savings an EV delivers over a comparable gas car, but they are not zero, and they land disproportionately on exactly the drivers — renters, apartment dwellers, high-mileage commuters without home charging — who are least able to simply shrug off an unreliable network.

Beyond the network: state and federal complaint channels

There is no single national hotline built for an individual driver to report a broken charger to the government, but there are real channels adjacent to the problem, and it's worth knowing what each one actually does before you reach for it.

The Joint Office of Energy and Transportation's ChargeX Consortium is the closest thing to a federal reliability program, but it works upstream of individual complaints — it's a DOE national-lab collaboration that develops industry-wide standards, not a complaint desk [24][25]. Its two most concrete deliverables to date were a best-practices report on payment reliability, grouping failure causes into network, integration, hardware, customer-experience and maintenance categories, and a formal set of 26 minimum required error codes meant to give every charger, car and network operator the same diagnostic language [26][28]. Neither one processes a report from you directly, but both exist because enough drivers hit the same failures for long enough that the federal government funded a fix.

NEVI's 97% uptime standard is real regulatory teeth, but it points at the funding recipient, not at your specific bad experience — a state DOT or its contracted operator can lose federal funding eligibility for chronic underperformance under 23 CFR Part 680, but there's no individual-driver claims process attached to that standard [6][7]. The Joint Office's parallel price-transparency work pushes a related but different issue: getting states to adopt uniform "legal metrology" standards, based on NIST Handbooks 44 and 130, so that a public charger's kWh meter is certified accurate the same way a gas pump or a grocery scale is [27][32].

That metrology angle is where a genuine state-level complaint channel exists. Several states run weights-and-measures programs that inspect public chargers for billing accuracy, and at least one has acted on driver complaints directly: Maryland's weights-and-measures inspectors began dispatching to charging stations in 2025 in response to customer complaints, and found — and temporarily shut down — chargers illegally billing by time instead of by the kilowatt-hour actually delivered, a practice that can penalize a driver simply for charging on a slower car [31]. California's AB 1423 similarly directs the state to build uptime recordkeeping and reporting standards with an administrative enforcement process for violations [33]. If your complaint is specifically about being billed for energy you didn't receive — not just a charger that wouldn't start — your state's weights-and-measures or consumer-protection office, not the network, is the channel built to hear it.

The FTC's ReportFraud.ftc.gov exists for consumer complaints broadly, including paying for a service that was never delivered, but it's worth being precise about what it does: the FTC states plainly that it cannot resolve an individual report, and instead uses the volume and pattern of reports to investigate and build cases against businesses [29][30]. It is a real, legitimate channel for a documented pattern of being charged and never receiving a service — keep it in reserve for a network that repeatedly refuses to refund a clearly failed, provably paid session — but it is not a fast path to getting your $12 back; the network's own support line, used promptly with evidence in hand, remains the more direct route for that.

Key findings

ChargeCostLab's research for this piece converges on five load-bearing conclusions:

• ChargeCostLab's review of J.D. Power's 2025 survey finds 14% of US public charging visits ended without a successful charge, an improvement from 19% in 2024, with 60% of failures caused by an out-of-service charger rather than driver error [1].
• ChargerHelp's 2025 dataset of over 100,000 sessions found first-time charge success at 71% even where operators self-report 98%+ uptime, and that success falls to 69.9% by a station's third year — evidence that "uptime" and "worked when I plugged in" are two different metrics [2][3].
• None of the five major US charging networks — Tesla, Electrify America, EVgo, ChargePoint or Blink — publishes an automatic refund guarantee for a failed session; EVgo's terms go further and state charges are non-refundable by default unless the company decides otherwise [12].
• ChargeCostLab's own calculation puts the annual cost of failed sessions, retries and detours at roughly $310 and 10.5 hours for a public-charging-dependent driver, against about $78 and 2.6 hours for a driver who charges mostly at home (our calculation).
• A genuine government enforcement channel exists for billing-accuracy complaints — state weights-and-measures programs such as Maryland's have inspected and shut down chargers billing incorrectly — but no federal agency currently resolves an individual driver's single failed-session dispute [31].

Frequently asked questions

The answers below are self-contained; full assumptions and every cited figure trace to the sources list.

Methodology note

Failure-rate and reliability figures are drawn from J.D. Power's 2025 EVX study, ChargerHelp's 2025 and 2024 reliability reports, a UC Berkeley field study and Paren's Q1 2026 index, each independently re-checked against the original or a verified mirror where the publisher's own page allowed automated access [1][2][3][4][5][8]. Every network reporting-channel and refund-policy claim is sourced to that network's own support, FAQ or terms page; where a page actively blocked automated retrieval, the claim is corroborated against a second source and noted in the text rather than presented as independently confirmed [9]–[20]. The annual cost-of-failure figures are ChargeCostLab's own calculation, built on the cited failure rates with explicitly stated time-value and detour-distance assumptions, and are reproducible by substituting your own driving pattern into the same formula.