The End of Leap Seconds: What the New Timekeeping Decision Actually Changes

The international timekeeping conference has approved continuous UTC from May 20, 2027, ending the routine leap-second mechanism. The decision behind today’s circulating headlines does not make seconds longer, remove leap years or require people to move their clocks forward by an hour.

Headlines saying the world has abolished leap seconds are based on a real decision, but the details matter. Resolution 5 of the 2026 General Conference on Weights and Measures sets May 20, 2027, as the effective date for continuous Coordinated Universal Time, or UTC. It also increases the permitted difference between UTC and Earth-rotation time to 3,600 seconds.

That one-hour tolerance is a limit on how far two time scales may diverge over a very long period. It is not an instruction to add an hour to your watch next May. Nor does it redefine the physical duration of a second. The reform changes the way civil time is kept related to an irregularly rotating Earth.

The subject is being discussed on social media as well as in specialist reporting. The official resolution is the evidence for the decision; social posts are evidence of discussion, not an authority on the technical rules. Here is what the development actually means.

The vote is real, and the date is not today

Timeanddate’s report, updated October 15, records the conference’s approval and explains the change through interviews with timekeeping specialists. It identifies the widening tolerance between UTC and UT1, the time scale based on Earth’s rotation. Its account corroborates the central outcome in the official resolution.

The report also addresses a tempting description: replacing leap seconds with leap hours. An IERS specialist explains that a possible adjustment far in the future remains a decision for future generations. The reform permits a much larger difference; it does not provide a scheduled leap-hour event for people alive today.

This is why a short headline can create the wrong impression even when its central claim is accurate. The conference has decided on a transition, with an implementation date. It has not suddenly accelerated the planet, changed the length of an ordinary workday or issued an instruction to reschedule tomorrow’s appointments. The scale of the permitted future difference and the date of the operational change answer different questions.

UTC comes from a worldwide measurement system

The BIPM’s explanation of UTC describes contributions from about 80 time laboratories and approximately 450 atomic clocks. The international reference is computed from clock and time-transfer data, while laboratories maintain local, real-time versions known as UTC(k).

The BIPM publishes comparisons that allow those laboratories to relate their local signals to the international reference. This is the infrastructure behind synchronization across telecommunications, navigation, science and other systems. The reference is therefore not simply the reading of one exceptionally accurate clock in one building.

The distinction helps explain why agreement matters. Two organisations can each have stable clocks and still have difficulty comparing events if their time references are handled differently. A shared system supplies the relationship between those readings. The reform concerns continuity within that shared reference, rather than an attempt to make every device an independent judge of Earth’s rotation or of the exact position of the Sun.

Why a clock ever displayed 23:59:60

NIST’s leap-second record explains the historical mechanism. An extra second can be inserted after 23:59:59, producing 23:59:60 before midnight. Its table records the first insertion on June 30, 1972, and the most recent on December 31, 2016.

The same record distinguishes UTC from International Atomic Time, or TAI. The accumulated difference is currently 37 seconds, which includes the initial relationship between the scales as well as subsequent adjustments. It should not be confused with the much smaller difference between UTC and UT1 that leap seconds were designed to control.

These labels can seem unnecessarily technical until a claim mixes them together. Saying atomic time is ahead by dozens of seconds does not mean a phone is dozens of seconds wrong. It means the two named scales assign different time labels according to their definitions. A meaningful comparison needs to identify both scales, rather than treating every number called an offset as the same measurement.

Earth’s rotation is not a perfectly uniform clock

NIST’s explanatory FAQ describes UTC as an atomic scale and UT1 as a scale based on the irregular rotation of Earth. Rotation can speed up or slow down over shorter intervals, even though the long-term tendency is toward slower rotation.

The FAQ also distinguishes leap seconds from leap years. Leap-second decisions respond to measurements of rotation, whereas calendar leap days follow rules that keep the year related to Earth’s orbit around the Sun. Ending one mechanism does not remove February 29 from the Gregorian calendar.

For a reader checking an alarming post, that difference is a useful first test. A claim that a timekeeping vote has cancelled leap years is combining two separate systems. Both relate human timekeeping to astronomical cycles, but they do so at different scales and through different rules. Similar names do not make them interchangeable, and changes to one should not be used to infer an unannounced change to the other.

A single extra second can complicate computer records

The NIST Internet Time Service documentation explains a concrete problem. Some digital representations cannot express the label 23:59:60. NIST’s described implementation repeats the binary time corresponding to 23:59:59, creating two consecutive seconds with the same label.

That can make it difficult to determine the order of events near the adjustment. A later event can carry a numerical timestamp that looks earlier if the comparison ignores the repeated second. Measuring an interval across the adjustment can create a related ambiguity.

This gives the reform a practical significance beyond the unusual appearance of a clock display. Software often needs to compare records from different machines. If those machines handle an exceptional second differently, a simple numerical comparison may not explain the real sequence. Removing routine discontinuities can simplify future operations, although historical records will still need to be interpreted using the rules that applied when they were created.

The system still needs observations of Earth

The IERS bulletin catalogue separates several information products. Bulletin A covers rapid Earth-orientation determinations, Bulletin B monthly orientation parameters, Bulletin C leap-second announcements and Bulletin D the broadcast value of DUT1, the difference between UT1 and UTC.

The existence of these separate products illustrates that observing Earth’s orientation is a broader scientific task than deciding whether to insert a second. A continuous civil reference does not make astronomical measurements unnecessary. Applications that need rotation-based information will still require the appropriate observations and conversions.

For someone following a satellite, telescope or geodetic measurement, the relevant question is which reference a calculation requires. The civil timestamp can identify when an observation occurred, while an orientation model supplies information about Earth at that moment. Keeping those functions distinct is more accurate than saying scientists have abandoned the Sun or stopped caring how fast the planet turns.

The reform follows years of technical preparation

The CCTF’s July 2026 preparatory report describes work with astronomical organisations, satellite-navigation providers, national measurement laboratories and telecommunications bodies. It traces the process back to the 2022 decision to increase the permitted UTC–UT1 difference in or before 2035.

That document is background on how the proposal developed, not the final adopted text. Distinguishing the two is important when older pages continue to describe a draft, while newer reporting discusses an agreed resolution. The latest official decision should determine the description of the outcome.

It also explains why implementation involves many organisations. Time information crosses systems designed at different dates for different purposes. An international decision creates a common direction, while operators still need to understand how their equipment and services represent and distribute time. A transition date is therefore part of an engineering process, rather than a moment when every clock owner must take the same manual action.

A local time signal is different from the international calculation

NIST’s account of its local UTC realization describes an ensemble of clocks continuously compared under controlled conditions. An algorithm combines their measurements, giving the resulting time scale advantages over relying on a single instrument. Coordination adjustments then relate the local realization to UTC.

This distinction between a reference calculation and the physical signals distributed by a laboratory is another reason the reform cannot be understood as everyone resetting one master clock. Timing services connect measurements, algorithms and transmission systems. A change in the way time labels are coordinated does not imply that the underlying oscillators stop or jump. Understanding which layer is changing makes the practical consequences clearer.

Redefining the second is a separate scientific project

The BIPM’s roadmap for redefining the second describes work toward a possible future definition reflecting improvements in atomic-clock technology. That work is separate from changing the leap-second mechanism. A proposal concerning the unit and a decision concerning the continuity of a time scale should not be collapsed into one event.

NIST’s account of measuring a second explains the present definition through a specified frequency associated with caesium atoms. It also describes advances in optical clocks. Better ways to realize a unit concern precision and reproducibility; they are not permission to arbitrarily stretch time.

For everyday readers, the central verified development is a coordinated change to a global reference system next year. The useful response to a dramatic headline is to check the adopted text, its effective date and the exact quantity being changed. Those checks reveal an important infrastructure reform without turning it into a fictional extra hour or a disappearing calendar day.

See our viral-story verification guide and more Smashology Viral reporting for source-linked checks of claims circulating online.