The leap second was created to keep human clocks aligned with Earth’s rotation. Now digital infrastructure β and a genuinely unprecedented risk β are pushing the world toward abandoning it.
For thousands of years, humans measured time by looking at the sky. The Sun’s movement determined the day, the shifting stars marked the seasons, and Earth’s rotation set the natural rhythm around which civilizations organized agriculture, navigation, and daily life.
Today, most of the world’s clocks run on a very different foundation. Modern timekeeping depends on atomic clocks, computers, satellites, and tightly synchronized networks β measuring time with a precision earlier generations couldn’t have imagined. But there’s a problem: Earth doesn’t rotate with perfect regularity. That small, persistent mismatch between the planet’s actual rotation and the world’s official clocks produced one of the stranger inventions in modern science, the leap second. And this October, timekeepers from around the world are set to vote on ending it for good.
Why Does a Leap Second Exist?
Coordinated Universal Time, or UTC, is the world’s principal reference time scale, built on extraordinarily precise atomic clocks. But civil time also needs to stay reasonably close to Earth’s actual rotation, a measure known as UT1 β and because Earth’s rotation isn’t perfectly steady, the two drift apart gradually. Since 1972, international authorities have inserted an extra second into UTC 27 separate times to keep that drift within an agreed limit, most recently on December 31, 2016.
The system worked well enough when clocks were mostly used by people glancing at a wall or a wrist. The digital world has made it a genuinely harder problem.
A Single Second Can Disrupt a Digital System
Modern infrastructure runs on synchronized clocks β telecommunications networks, satellite navigation, electricity grids, financial systems that timestamp every transaction, data centers coordinating enormous numbers of processes. For a person, an extra second is meaningless. For a computer system built around an unbroken sequence of timestamps, it can cause real trouble; leap seconds have been directly linked to documented outages at companies including Reddit, Cloudflare, and Qantas over the years. Some companies now handle it through “smearing” β spreading the extra second gradually across a longer window rather than inserting it all at once β but different companies smear differently, which means the fix meant to keep systems synchronized can itself introduce new discrepancies between them.
Meta has been especially vocal about this problem, publicly advocating for the leap second’s elimination after experiencing its own infrastructure headaches from past insertions.
The Problem Nobody Has Tried Before
What’s actually accelerating this decision is something new: Earth’s rotation has been speeding up. The planet set rotation-speed records in both 2024 and 2025, and current projections put the odds at roughly 30 percent that the world will need its first-ever negative leap second β clocks skipping a second rather than adding one β sometime before 2035. That’s never happened, and there’s no software patch history, no operational precedent, and no confident way to predict how every interconnected digital system would handle an event many of them were never designed for in the first place. The worry isn’t that computers everywhere would suddenly fail. It’s that nobody can say with real confidence how they’d behave.
The World Is Preparing to Change the Rules
The move away from leap seconds isn’t new β the CGPM’s member states agreed back in November 2022 to increase the allowed gap between UTC and Earth-rotation time by 2035 at the latest, explicitly to move toward a continuous UTC and cut the risk to critical infrastructure. But the negative-leap-second risk has compressed that timeline considerably. The 28th General Conference on Weights and Measures meets in Versailles, France, from October 13 to 15, 2026, and member states will vote on a resolution to replace the leap second with something far rarer: a leap hour, applied only once the gap between atomic and astronomical time reaches a full 3,600 seconds. If approved, the mechanism could be retired as early as 2027, with the century-old link between UTC and a one-second tolerance gone for good.
There’s genuine diplomatic friction behind the timing, too. Russia previously lobbied for the more distant 2035 deadline specifically to give its GLONASS satellite-navigation system β which, unlike GPS, still incorporates leap seconds directly β time to adapt, and that same tension may resurface as engineers push back on an accelerated 2027 rollout.
Why an Hour?
Letting clocks drift by a full hour sounds extreme at first, but the goal is precisely to make corrections so rare that digital systems essentially never have to deal with them. At Earth’s current rate of rotational variation, reaching a full hour of accumulated drift would take an estimated 50 to 100 years β effectively removing the leap-second problem from the engineering agenda for the foreseeable future, and possibly for the rest of this century. Felicitas Arias, former director of the timekeeping department at the Bureau International des Poids et Mesures, has said she expects the International Telecommunication Union to support the Versailles vote, though the ITU’s formal cooperation isn’t fully locked in.
The philosophy is straightforward: instead of repeatedly nudging the world’s clocks by one second, let astronomical time and atomic time slowly drift apart, and let the correction β if it’s ever needed β happen once, on a scale that’s simple to plan around. People wouldn’t notice any difference in daily life. Digital systems would gain something far more valuable to them: predictability.
The Hidden Cultural Cost
From an engineering standpoint, the case is compelling. But time was never purely an engineering problem. There’s something genuinely meaningful in the fact that the clock on your phone is, however indirectly, tied to the movement of the planet under your feet. For most of human history, time was astronomical β the day from Earth’s rotation, the month from the Moon, the year from Earth’s orbit. Atomic time represented a dramatic break from that, defining the second through the behavior of atoms rather than the movement of anything celestial. Abolishing the leap second is another step in that same direction: our official clocks would no longer need to track Earth’s actual rotation at all, at least not on any timescale a human being would ever personally experience.
BIPM officials have been careful to note that the underlying relationship doesn’t disappear entirely β as one framing put it, UTC “remains relevant to the Earth,” even as the two measures are allowed to separate for far longer stretches than before.
Atomic Clocks Are Becoming Extraordinary
There’s a real irony in what’s driving this shift: the technology making it necessary is also one of humanity’s most impressive scientific achievements. Researchers are now developing optical atomic clocks β built on elements like ytterbium β that reach levels of precision far beyond today’s already remarkable standards. Ancient civilizations looked to the sky because the stars and Sun were the most reliable clocks available to them. Modern civilization has built clocks so precise that Earth’s own rotation has effectively become the less reliable reference point. Earth, in other words, is no longer the best clock we have.
The Digital Economy Needs Predictable Time
The leap-second debate says something bigger about modern infrastructure: society increasingly runs on invisible systems that coordinate not just information, but the precise timing of information. A telecommunications network doesn’t just transmit data β it sequences it in time. A satellite-navigation system calculates position partly through extraordinarily precise timing signals. Financial networks don’t just record trades β they establish the exact order in which they happened, which can matter enormously in high-frequency markets. Energy grids coordinate generation and distribution across huge geographic areas in real time. In every one of these systems, time isn’t a backdrop β it’s infrastructure in its own right, which is exactly why what once looked like a harmless astronomical footnote has become a serious engineering concern.
The Risk of Making Rare Events Even Rarer
Not everyone thinks abolishing the leap second is an unambiguous improvement. One real objection: making an event extremely rare can make it harder to handle competently when it finally arrives. Engineers stay sharp on systems they test regularly; a correction that happens once every 50 to 100 years can’t realistically be rehearsed the way an occasional one-second adjustment can be. That trades a small, recurring inconvenience for a much larger, much more distant one β and there’s a case that the distant version is actually the riskier one to get wrong, precisely because nobody currently working in the field will have direct experience managing it. For today’s infrastructure, though, the trade-off still looks like the better bet: a continuous timescale is simply easier to design systems around than one that occasionally introduces an unexpected jump.
This Is Bigger Than a Clock
The leap-second debate might sound trivial next to wars, elections, or AI, but it illustrates something genuinely larger: modern civilization increasingly builds its infrastructure around what machines need rather than what nature happens to provide. Computers want consistency. Networks want continuity. Algorithms want predictable inputs. Earth’s rotation doesn’t naturally cooperate β it fluctuates as the atmosphere shifts and the planet’s interior moves, and the length of a day isn’t perfectly constant. Human technology has reached a point where those tiny variations genuinely matter, and the push to separate UTC from Earth’s rotation more permanently is really civilization adapting the natural world’s rhythms to the requirements of digital infrastructure, rather than the other way around.
What Will People Actually Notice?
Almost nothing. If the leap-hour proposal is adopted, people will keep using hours, minutes, and seconds exactly as they do now. Midnight stays midnight. Nobody experiences a missing hour in their lifetime, and quite possibly nobody alive today ever will. The change lives almost entirely beneath the surface, inside metrology standards and the technical systems that keep modern civilization synchronized. Its significance is symbolic as much as practical β a small, deliberate loosening of the tie between our clocks and the planet’s rotation, decided in a conference room in Versailles rather than felt by anyone going about an ordinary day.
The End of an Ancient Relationship
Clocks were invented to organize human life, but for most of history they stayed tethered to the heavens. The leap second was an unusual compromise between two worlds β ancient astronomical timekeeping and modern atomic precision. Now, driven less by philosophy than by a genuinely novel engineering risk, that compromise appears headed for retirement: the digital world wants a perfectly continuous clock, Earth’s rotation refuses to provide one on cue, and this October, in Versailles, humanity looks likely to formally choose the machine’s version of time over the planet’s.
That may well be the sensible engineering call β and it may also be close to inevitable for a civilization this dependent on precise, interconnected machines. But if the vote goes through and the leap second quietly disappears sometime after 2027, something more than a technical standard will have shifted. For the first time in human history, the world’s official clock will be deliberately allowed to drift, by design, away from the rotation of the planet that gave us the very idea of a day in the first place.
Author
