The end of leap seconds marks a rare moment when a tiny technical fix in global timekeeping becomes a major infrastructure decision. For most people, the change will be invisible. For engineers, astronomers, telecom operators, and anyone running distributed software, leap second elimination changes the rules for how Coordinated Universal Time is maintained, how clocks are synchronized, and how systems handle the slow drift between atomic time and the Earth’s rotation.
The debate matters because modern society depends on time in two different ways at once. Civil time needs to be predictable for databases, log files, legal records, and networked systems. Astronomical time needs to stay tied to the sky, because the length of the day is not perfectly constant. The old compromise was to patch Universal Time with occasional leap seconds. The new compromise is to stop patching and let the clock drift, at least for a very long time.
What a leap second was actually doing
A leap second was a one-second adjustment added to UTC to keep it from drifting too far from Earth rotation time. The underlying second comes from the International System of Units, not from the spinning planet. That means the modern second is defined by atomic physics and realized by highly stable atomic clocks, while the day is defined by astronomy and geophysics. Because Earth’s rotation is irregular, the two do not stay aligned forever.
Historically, civil time borrowed much of its rhythm from the sun. Greenwich Mean Time and later UTC helped standardize that relationship. But UTC became a compromise between the precision of atomic time and the messiness of planetary motion. The result was the leap second, usually inserted at the end of June or December, when the difference between UTC and Earth-based rotation time approached a threshold set by international standard-setters.
The institutions behind that compromise matter. The International Earth Rotation and Reference Systems Service monitors Earth orientation, the International Telecommunication Union handles global communications standards, and the International Bureau of Weights and Measures plays a central role in the international time scale. Their work sits behind something ordinary users rarely notice: every stable timestamp they see on a phone or server depends on an enormous amount of hidden coordination.
Since 1972, UTC has accumulated 27 positive leap seconds. There has never been a negative leap second, even though the concept has long existed in the standards discussion. Today, UTC is 37 seconds behind International Atomic Time (TAI). That number sounds tiny, but it is a useful reminder that
Frequently Asked Questions
If leap seconds were only one second long, why were they such a problem for software and networks?
Because many systems assume time moves forward continuously and monotonically. A leap second breaks that assumption: clocks can repeat, pause, or behave inconsistently for a moment. That can confuse databases, log ordering, distributed systems, authentication protocols, and telecom equipment, even if the physical difference is tiny.
Does eliminating leap seconds mean UTC will stop matching the Sun and the seasons?
Not immediately, but over very long periods, yes. Without leap seconds, UTC will slowly drift away from Earth rotation time as the planet’s spin changes. The difference grows gradually, so civil time stays practical for generations, but the sky-based definition of noon will become less exact over centuries.
What happens to the gap between UTC and atomic time now that leap seconds are being removed?
The gap will keep increasing steadily instead of being corrected intermittently. UTC is already 37 seconds behind TAI, and that offset will grow as the years pass. This makes UTC easier for atomic-clock-based systems to maintain, but it also means civil time becomes less tied to astronomical time.
Will ordinary users notice any change on phones, computers, or the internet?
Most users will notice nothing day to day. The main benefit is behind the scenes: fewer clock glitches, fewer special-case fixes, and less risk of timestamp errors in cloud services, logs, and global networks. The change matters most to engineers who maintain time-dependent infrastructure, not to casual users.
Why not keep leap seconds and just make devices handle them better?
In theory, that sounds simpler than changing UTC. In practice, it is hard to guarantee consistent behavior across billions of devices, operating systems, routers, and legacy systems. A single global standard that avoids leap seconds reduces complexity far more reliably than asking every implementation to handle rare edge cases perfectly.