guides
Public EV Charging Reliability: Uptime vs Failed Visits vs First-Time Success
Last verified October 27, 2026
Re-verification overdueThis guide remains available for context but is not presented as current until it is rechecked.
Update, October 26, 2026: A prior version described the California Energy Commission's September 24 hearing as upcoming. That hearing date has passed, and CEC now hosts a presentation and recording. The current program overview still labels the proposed Megawatt Charging System (MCS) changes proposed; the inspected docket does not itself establish a final adoption decision. Existing covered-charger uptime and reporting requirements remain separately described below.123
There is no single percentage in the current sources below that answers every version of “how often do public EV chargers fail?” The most useful U.S. reliability numbers measure different things.
J.D. Power's August 2026 study release says 12% of public-charging visits in its most recent quarter ended without the owner being able to charge, down from 14% in the same period a year earlier.4 ChargerHelp's 2025 analysis instead says 71% of charging attempts succeeded on the first try in its dataset.5 Federal rules use a third measure: covered NEVI/Title 23-funded charging ports must maintain greater than 97% average annual uptime, using a specific port-time formula and exclusions.67 California now adds a separate state compliance metric: a defined cohort of publicly funded DC fast chargers must meet a 97% per-port uptime standard for six years, with its own scope, exclusions and reporting rules.18
Those figures are not mathematically interchangeable. A charger can be counted as available for most of the year and still fail a first attempt. A driver can fail on the first attempt but succeed after retrying. A public-charging visit can also end without any charge even though the underlying problem was not simply “this charger was broken.” Regulatory thresholds are requirements for covered infrastructure, not observed national or statewide driver-success rates.
The current U.S. reliability numbers, side by side
| Metric | Current published figure | What is being measured | Scope / period | What it tells a driver | What it does not tell you |
|---|---|---|---|---|---|
| J.D. Power non-charge visit rate | 12% in the most recent quarter of the 2026 release, down from 14% in the same period a year earlier.4 | A driver visit to a public charging station where the owner was unable to charge. | 2026 EVX Public Charging Study; 6,594 BEV/PHEV owners; study fielded Jan.-Jun. 2026 and analyzed quarterly; public Level 2 and DC fast charging.4 | How often surveyed owners reached a public station but did not get a charge during the measured visit outcome. | It is not the percentage of physical chargers that are broken, and it is not a first-attempt success rate. |
| ChargerHelp first-time charge success rate (FTCSR) | 71% of attempts succeeded in the 2025 analysis.5 | Whether charging successfully starts on the first attempt. | ChargerHelp 2025 report summary; more than 100,000 sessions across 2,400 chargers.5 | How often a charging attempt works without a retry/reset/error path in that analyzed dataset. | A first-attempt failure does not necessarily mean the entire visit ended without a charge. It is not the same denominator as J.D. Power's visit outcome. |
| Federal minimum uptime | Greater than 97% average annual uptime per covered port.6 | Port time: hardware/software online and available or in use, plus successful electricity dispensing at the required minimum power, using the regulation's formula.6 | NEVI Formula Program projects and specified publicly accessible Title 23-funded charging projects; calculated monthly over the previous 12 months.67 | The federal availability/performance requirement for the ports covered by Part 680. | It is not an observed nationwide success rate for all public chargers and does not mean 97% of driver attempts succeed on the first try. |
| California covered-DCFC uptime | 97% uptime per covered port.18 | Port-time availability/performance under California's EV Charger Data and Reliability Standards. CEC says the charger hardware/software must be online, the port available for use, and electricity successfully dispensed.8 | Defined publicly funded DCFC cohort installed on or after Jan. 1, 2024, subject to stated exceptions; six years from installation.18 | California's state performance requirement for the covered charger cohort, plus a separate reporting framework. | It is not an observed statewide driver-success rate, does not apply to every California charger, and is not textually the same threshold as the federal greater than 97% requirement. |
| ChargerHelp outage recovery | 36% of outages need repeat site visits, causing 48% of total downtime in its 2026 report.9 | Repair/recovery burden after an outage. | ChargerHelp 2026 Annual Reliability Report / operational recovery dataset.9 | How repeat repair visits can concentrate downtime after a charger is already in an outage state. | It is not the percentage of chargers that fail or the probability that a driver will encounter a failed session. |
Why 97% uptime can coexist with a much lower first-time success rate
Uptime is fundamentally a time metric. First-time charge success is an attempt metric. Non-charge visits are a visit outcome metric.
The federal rule illustrates why the distinction matters. A covered port is considered up when its hardware and software are online and available for use, or in use, and the port successfully dispenses electricity at the required minimum power.6 The rule calculates uptime over a rolling 12-month period and excludes specified outage minutes caused by utility interruption, vehicle faults, scheduled maintenance, vandalism, natural disasters, and hours outside the station's identified operating hours.6
California's state standard is also a time/per-port compliance measure, but it is a different rule with a different covered cohort and its own exclusions/reporting process.18 That is why even two requirements centered on 97% uptime should not be collapsed into one undifferentiated national statistic.
These definitions answer compliance/availability questions. They do not say that every vehicle, payment path, connector interaction, authentication event, or first attempt will succeed.
ChargerHelp's 2025 report summary makes the attempt-level distinction explicit. It reports 71% FTCSR while citing reported uptime in the 98.7%-99.9% range for the charging context it analyzed.5 ChargerHelp defines FTCSR around whether a driver can successfully start charging on the first attempt.5
The numbers can therefore coexist without contradiction: uptime measures time-based availability under a defined methodology, while FTCSR asks what happened when a charging attempt was made and J.D. Power's metric asks whether a visit ultimately produced a charge.
What J.D. Power's 12% actually means
J.D. Power defines a non-charging visit as an instance where an EV owner arrives at a public charging station but is unable to charge the vehicle.4
Its August 12, 2026 release says the industry non-charge-visit rate reached 12% in the most recent quarter, the lowest level in the study's history, compared with 14% during the same period a year earlier.4
The time wording matters. Although the 2026 study was fielded from January through June, the 12% figure is described in the release as the most recent quarter. It should not be relabeled as a single first-half 2026 failure rate.4
The prior 2025 study provides more detail about failed visits. It reported 14% of owners visiting a public charger without successfully charging, and said the most common reason among those failed visits was an out-of-service or improperly working charger, cited in 60% of failed charging visits.10
That reinforces another distinction: non-charge visit includes the whole driver outcome. It should not be rewritten as 12% of chargers are broken.
What ChargerHelp's 71% actually means
ChargerHelp's September 2025 report summary says its analysis included more than 100,000 sessions across 2,400 chargers and reported a 71% first-time charge success rate.5
It describes FTCSR as whether the driver can successfully start charging on the first attempt.5 The same source says new stations averaged 85% success, while stations at year three averaged 69.9%.5
That makes FTCSR useful for a different question: Did the session start cleanly the first time?
It does not establish that 29% of public-charging visits ended without any charge. A failed first attempt may be followed by a successful retry, another port, a reset, or another recovery path. The public source used here does not provide a valid conversion from FTCSR to J.D. Power's non-charge-visit metric.
What the federal greater-than-97% rule actually means
Current 23 CFR 680.116 says states or other direct recipients must ensure each covered charging port has average annual uptime greater than 97%.6
The requirement is calculated monthly for the prior 12 months. Its formula subtracts specified excluded outage minutes, including outages caused by electric utility service interruptions, vehicle faults, scheduled maintenance, vandalism, natural disasters, and hours outside identified operation.6
The scope is also narrower than every public charger in America. Section 680.102 applies Part 680 to NEVI Formula Program projects and specified publicly accessible EV charging projects funded under Title 23.7
So greater than 97% is a regulatory requirement for covered projects, not a measurement proving that all U.S. public chargers operate above 97% and not a promise that 97% of first attempts will succeed.
California's 97% rule is similar, but it is not the federal rule
The California Energy Commission's current compliance guidance says its uptime/performance standard applies to a defined group of DC fast chargers that received state, ratepayer, or federal funding passed through a state agency, were installed on or after Jan. 1, 2024, and do not fall into the listed exclusions.18 Covered chargers must comply for six years from installation.8
CEC states the performance standard as 97% uptime per port.8 Its public guidance says uptime requires the DCFC hardware and software to be online, the charger to be available for use, and electricity to be successfully dispensed. The current page lists excluded downtime for vandalism up to 10 days per incident, grid power loss, natural disasters, and planned maintenance up to 72 hours per year.8
California also attaches a reporting transition to the install date. Under current CEC guidance:
- applicable DCFCs installed by Sept. 27, 2026 use the semiannual uptime-reporting path, with the first report due Jan. 31, 2027 for July 1-Dec. 31, 2026; and
- covered new DCFCs installed after Sept. 27, 2026 must support OCPP 2.0.1; networked units on that path submit hourly reliability data through the CEC's data-transfer process.18
That protocol/reporting requirement still does not turn 97% into a driver-experience statistic. It tells operators and regulators how a covered California cohort is measured and reported.
CEC held its Sept. 24, 2026 public hearing on proposed amendments to partially exempt Megawatt Charging System chargers from certain requirements and make minor refinements. Its meeting page now includes a hearing presentation and recording.2 As checked October 26, CEC's current program overview still describes the MCS amendments as Proposed Changes. Docket 22-EVI-04 lists the September 24 hearing presentation and September 25 comments; that inspected docket does not establish adoption of the proposed amendments, but the absence of a listed adoption order is not proof of the outcome of every possible regulatory process.13 This guide therefore does not treat the proposal as effective law.
Reliability after a charger goes down: recovery is another metric
ChargerHelp's current 2026 report moves the focus from first-time session success toward outage recovery and Mean Time To Resolution. Its public report page says 36% of outages need repeat site visits and those outages cause 48% of total downtime.9
That is useful because two networks could experience the same number of outages but recover from them at very different speeds. It is still a different denominator from driver visits, session attempts, or annual port uptime.
Do not read 36% as a charging-failure rate. It is the share of outages in ChargerHelp's report that require repeat site visits.
A quick test before comparing two charger-reliability percentages
Before putting two reliability numbers side by side, check five things:
- Unit: Is it measuring time, ports, attempts, sessions, visits, or outages?
- Outcome: Does
failuremean no first-try start, no charge at all, an offline state, or unresolved repair? - Population: Is it a survey of drivers, an operator/maintenance dataset, a federal funding class, a state funding class, or the whole public network?
- Period: Is it a quarter, a six-month fielding period, a rolling 12 months, a six-year compliance obligation, or a report-specific dataset?
- Exclusions: Does the metric exclude utility/grid outages, vehicle faults, maintenance, vandalism, natural disasters, closed hours, or other causes?
If those fields differ, the percentages should not be averaged or converted unless the originating source provides a valid bridge.
What not to conclude from these numbers
- Do not say
12% of public chargers are broken. J.D. Power's 12% is a driver non-charge-visit outcome for the most recent quarter in its 2026 release.4 - Do not say
29% of charging visits end in failure. ChargerHelp's 71% measures first-time charge success in its 2025 analyzed sessions.5 - Do not say
federal chargers work for 97% of drivers. The federal requirement is a port-time uptime standard with a defined formula, exclusions, and funding scope.67 - Do not say
California chargers work 97% of the time statewide. California's 97% figure is a performance requirement for a defined covered DCFC cohort, not an observed statewide census or attempt-success rate.18 - Do not say California's proposed MCS exemption is already effective. Although the Sept. 24 hearing has passed, CEC's program overview still labels the amendments proposed as of the October 26 check; the inspected docket does not establish adoption.123
- Do not say
36% of chargers fail. ChargerHelp's 36% figure is the share of outages that require repeat site visits in its 2026 recovery analysis.9 - Do not average 12%, 71%, the federal >97% threshold, California's 97% threshold, and 36% into one reliability score. The denominators and legal scopes are different.
Current trend that is directly comparable
The cleanest current year-over-year comparison in the checked sources is J.D. Power's own like-for-like series: 12% non-charge visits in the most recent quarter of the 2026 release versus 14% in the same period a year earlier.4
The ChargerHelp 2025 FTCSR figure and its 2026 outage-recovery statistics should remain separate. The current checked ChargerHelp 2026 public report surface does not provide a directly comparable replacement FTCSR figure that would support a 2025-to-2026 FTCSR trend on this page.59
Federal and California uptime figures on this page are regulatory requirements, not observed trend points. They should not be inserted into a driver-outcome trend line unless a future source publishes measured compliance using a comparable methodology.
When this page will be rechecked
This reference should be reverified when J.D. Power publishes a new EVX Public Charging update, ChargerHelp publishes/revises a reliability report, the Joint Office/ChargeX ecosystem materially changes its reliability metric guidance, the federal Part 680 uptime/applicability language changes, or California changes its EV Charger Data and Reliability Standards. The September 24 CEC hearing and September 28 reporting-installation boundary have already passed. Future California triggers are a posted final/adopted MCS order or new express terms, changes to the current rule/implementation instructions, and the January 31, 2027 first semiannual reporting deadline for the applicable earlier-installation cohort.182 Recheck by November 2, 2026 or sooner on an official change.
Sources
Footnotes
-
California Energy Commission, EV Charger Data and Reliability Standards — current program overview checked Oct. 26, 2026. https://www.energy.ca.gov/programs-and-topics/programs/clean-transportation-program/ev-charger-data-and-reliability-standards Establishes the current simplified applicability summary, 97% covered-DCFC uptime requirement, Sept. 28 OCPP/hourly-reporting transition, data-sharing/enrollment context, and the pending proposed MCS changes. CEC directs users to full regulatory text for controlling detail. ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11
-
California Energy Commission, Public Hearing — Notice of Proposed Action: Rulemaking to Amend Electric Vehicle Charger Reliability Standards — current hearing page checked Oct. 26, 2026; Sept. 24, 2026 hearing now documented with posted presentation and recording. https://www.energy.ca.gov/event/meeting/2026-09/public-hearing-notice-proposed-action-rulemaking-amend-electric-vehicle Establishes that CEC proposes amendments to partially exempt MCS chargers from certain requirements and make minor refinements; it does not establish adoption. ↩ ↩2 ↩3 ↩4
-
California Energy Commission, Docket 22-EVI-04 — Electric Vehicle Charging Infrastructure Reliability — current docket checked Oct. 26, 2026. https://efiling.energy.ca.gov/Lists/DocketLog.aspx?docketnumber=22-EVI-04 Establishes the proceeding chronology, including Aug. 7, 2026 filing of proposed express terms/notice and the absence in the checked visible docket of a later adoption document. The docket chronology does not by itself establish every substantive provision. ↩ ↩2 ↩3
-
J.D. Power, 2026 U.S. Electric Vehicle Experience (EVX) Public Charging Study release — J.D. Power via Business Wire, Aug. 12, 2026. https://www.businesswire.com/news/home/20260812737541/en/Public-EV-Charging-Improves-as-OEM-Backed-Networks-Gain-Traction-JD-Power-Finds Establishes the non-charge-visit definition, 12% most-recent-quarter result, 14% same-period prior-year comparison, sample, fielding period, quarterly analysis, and public Level 2/DC fast scope. ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8
-
ChargerHelp, 2025 EV Charging Reliability Report summary — Sept. 24, 2025. https://www.chargerhelp.com/posts/chargerhelp-report-reveals-charge-success-rate-and-not-uptime-more-accurate-metric-for-ev-driver-experience.html Establishes the >100,000-session / 2,400-charger analysis, 71% FTCSR, FTCSR definition/framing, reported 98.7%-99.9% uptime comparison, and station-age comparison. ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10
-
23 CFR 680.116 — Minimum uptime and public charging information requirements — current eCFR checked Oct. 26, 2026. https://www.ecfr.gov/current/title-23/chapter-I/subchapter-G/part-680/section-680.116 Establishes the >97% covered-port annual uptime requirement,
updefinition, rolling calculation, formula, and exclusions. eCFR is authoritative but unofficial. ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 -
23 CFR 680.102 — Applicability — current eCFR checked Oct. 26, 2026. https://www.ecfr.gov/current/title-23/chapter-I/subchapter-G/part-680/section-680.102 Establishes the NEVI Formula Program / applicable Title 23-funded project scope for Part 680. ↩ ↩2 ↩3 ↩4
-
California Energy Commission, EV Charger Performance Standards and Uptime Reporting Regulations — current compliance guidance checked Oct. 26, 2026. https://www.energy.ca.gov/programs-and-topics/programs/clean-transportation-program/ev-charger-data-and-reliability-1 Establishes the covered DCFC cohort, six-year compliance duration, 97% per-port standard, public uptime definition, listed exclusions, OCPP 2.0.1 installation boundary, semiannual/hourly reporting paths, and Jan. 31, 2027 first semiannual deadline. This is simplified CEC guidance rather than a substitute for the full Title 20 text. ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12
-
ChargerHelp, 2026 Annual Reliability Report — current report surface checked Oct. 26, 2026. https://www.chargerhelp.com/2026-reliability-report.html Establishes the recovery/MTTR focus and the finding that 36% of outages need repeat site visits and cause 48% of total downtime. This page does not use the report landing page's ambiguous large data-volume label as a session-count claim. ↩ ↩2 ↩3 ↩4 ↩5
-
J.D. Power, 2025 U.S. Electric Vehicle Experience (EVX) Public Charging Study — Aug. 13, 2025. https://www.jdpower.com/business/press-releases/2025-us-electric-vehicle-experience-evx-public-charging-study Establishes the 2025 14% non-charge-visit result, failed-visit reason detail, sample, and fielding period. ↩
