Why Leap Seconds Are Being Abolished by 2035 (and What It Breaks)

Every few years, engineers around the world hold their breath as clocks tick from 23:59:59 to 23:59:60 before reaching midnight. That extra second — the leap second — has been quietly breaking software systems since 1972. In 2022, the International Bureau of Weights and Measures voted to abolish leap seconds by 2035, and the global tech community collectively exhaled.

I have spent years following time synchronization standards, and the leap second has always been one of those topics that sounds trivial until a major website goes offline because of it. When the leap seconds abolished 2035 decision was announced at the General Conference on Weights and Measures in Versailles, it resolved a decades-long battle between astronomers who rely on Earth-based time and software engineers who depend on atomic precision.

This article breaks down what leap seconds are, why they were created, the specific systems they have broken over the years, and what happens after 2035. Whether you work in network infrastructure or simply find timekeeping fascinating, understanding this shift matters because it changes how every digital system on Earth keeps time.

We will also cover the looming threat of a negative leap second — something that has never been tested in live systems and could cause significant disruption before 2035 arrives.

What Is a Leap Second?

A leap second is a one-second adjustment added to Coordinated Universal Time (UTC) to keep it synchronized with Earth’s irregular rotation. When our planet’s spin slows down slightly, atomic clocks and astronomical time drift apart. The leap second bridges that gap.

Leap seconds get inserted at the end of June 30 or December 31, creating a minute with 61 seconds instead of the usual 60. The timestamp sequence looks like this: 23:59:59, then 23:59:60, then 00:00:00. That 23:59:60 value is where the trouble starts for software that has never been told a minute can contain 60 seconds.

The International Earth Rotation and Reference Systems Service (IERS) monitors Earth’s spin and announces leap seconds roughly six months before they take effect. There is no fixed schedule — leap seconds happen only when Earth’s rotation has drifted far enough to warrant one.

Since the system began in 1972, 27 leap seconds have been added to UTC. All 27 have been positive, meaning an extra second was inserted into the clock. A negative leap second — removing a second instead of adding one — has never occurred but is becoming increasingly likely as Earth’s rotation accelerates.

Why Leap Seconds Exist: Earth Rotation vs Atomic Time

Two timekeeping systems run in parallel, and they do not agree perfectly with each other.

The first is International Atomic Time (TAI), based on the SI second defined by cesium atom oscillations. Atomic clocks are extraordinarily precise — they would not lose a full second in 30 million years. This is the time standard that computers, satellites, and communication networks prefer because it never jumps, skips, or reverses.

The second is Universal Time (UT1), based on the actual rotation of Earth relative to distant stars. Our planet is not a reliable timekeeper. Tidal friction from the Moon gradually slows Earth’s spin, while movements in the molten core, ocean currents, and even weather patterns cause unpredictable fluctuations in rotation speed.

Without correction, atomic time and astronomical time would slowly diverge. Over thousands of years, the Sun would eventually appear overhead at what clocks call midnight. Leap seconds keep UTC — which serves as the practical compromise between these two systems — within 0.9 seconds of UT1.

The complication is that Earth’s rotation has been accelerating since around 2020 rather than slowing. This raises the real possibility of needing the first negative leap second in history, something no production system has ever handled.

History of Leap Seconds: From 1972 to Today

Leap seconds were introduced in 1972 as a compromise between scientists who wanted atomic precision and astronomers who needed Earth-based time. Before 1972, the world used a system of fractional frequency offsets and rubber seconds that stretched or compressed the length of each second. That approach proved too complicated to maintain, and the clean leap second system replaced it.

The first leap second was added on June 30, 1972. Between 1972 and 1979, nine leap seconds were inserted — nearly one per year — because Earth’s rotation was slowing at a relatively steady rate. The pace slowed in the 1980s and 1990s, with gaps of a year or more between adjustments.

The most recent leap second was added on December 31, 2016. As of 2026, nearly a full decade has passed without a new one. This unusually long gap reflects a surprising acceleration in Earth’s rotation that scientists did not predict.

No leap second has been needed since 2016 because Earth has been spinning slightly faster than expected. Researchers attribute this to changes in the Chandler wobble, tidal effects, and possibly climate-related mass redistribution as polar ice melts and shifts weight toward the equator.

This unexpected acceleration is precisely why a negative leap second — subtracting time instead of adding it — is now a serious concern among timekeeping professionals.

Why Leap Seconds Are Being Abolished by 2035

The leap seconds abolished 2035 decision was driven by one straightforward reality: software systems cannot reliably handle a minute that contains 61 seconds.

On November 18, 2022, member states of the General Conference on Weights and Measures (CGPM) meeting in Versailles, France, voted to discontinue leap seconds no later than 2035. The resolution passed with overwhelming support from the global scientific and technical communities.

Patrizia Tavella, head of the BIPM Time Department, noted at the time that introducing leap seconds creates discontinuities that risk causing serious malfunctions in critical digital infrastructure. Every time a leap second is added, engineers across the world must manually prepare systems, apply patches, and hope nothing breaks during the transition.

The root problem is that computer software almost universally assumes time always moves forward at a constant, predictable rate. A 61-second minute violates that core assumption. Some systems freeze at the unknown 23:59:60 timestamp. Others repeat 23:59:59 twice. A few crash outright when they encounter a time value they cannot parse.

Major tech companies including Google, Amazon, and Microsoft have already developed their own workaround called the “smear” technique — gradually spreading the extra second across a 24-hour period so no single minute ever contains 61 seconds. But smear implementations from different vendors are incompatible with each other and with the official UTC standard, creating a fragmented and confusing timekeeping landscape.

What Leap Seconds Break: Real-World Failures

The theoretical problems with leap seconds become very real when production systems fail. Here are the most significant documented incidents.

The 2012 Reddit Outage

On June 30, 2012, a leap second triggered a cascade of failures across the internet. Reddit experienced a massive outage that lasted several hours. The root cause was a Linux kernel bug that caused Java applications to enter tight CPU-spin loops when the system clock jumped forward by one second.

Foursquare, LinkedIn, Mozilla, and StumbleUpon all reported significant technical issues during the same event. Many affected systems required manual restarts to recover, and some engineers spent their entire weekend tracking down time-related bugs.

The 2017 Cloudflare DNS Incident

When a leap second was added on December 31, 2016, Cloudflare’s DNS resolver experienced a partial outage that carried into January 1, 2017. The company’s RRDNS software assumed that timestamps were always monotonically increasing. The leap second caused the code to compute negative durations, which produced invalid DNS responses.

Some DNS queries returned garbage data for several hours until engineers deployed a fix. Cloudflare later published a detailed post-mortem explaining exactly how the assumption of monotonic time broke their system.

Airlines and Satellite Navigation

Aviation systems depend on precise UTC synchronization for flight scheduling, air traffic control, and route management. Different subsystems handle the leap second differently — some freeze, some skip, some smear. When adjacent systems disagree about what time it is during that critical second, scheduling conflicts and navigation discrepancies can result.

GPS sidesteps the issue by running on its own continuous time scale, GPS Time, which does not include leap seconds. Receivers then apply a known offset to convert GPS Time to UTC. This approach works but means the offset changes with every leap second, requiring receivers to download updated broadcast data.

Financial Trading Systems

High-frequency trading systems that execute transactions in microseconds face the most acute risk. A clock that jumps backward — or even appears to repeat a second — can cause transaction ordering errors. In the financial world, that means lost money, disputed trades, and potential regulatory violations.

Major exchanges typically halt trading around leap second events as a precaution. The cost of this downtime runs into millions of dollars per incident, all to prevent a single second from corrupting transaction logs.

Telecommunications Networks

Cellular networks synchronize their base stations using dedicated time servers. When leap seconds disrupt this synchronization, calls can drop and data connections can fail. Different equipment vendors implement leap second handling differently, which creates subtle inconsistencies between hardware from different manufacturers on the same network.

The root frustration is that there is no single correct way to handle a 61-second minute. POSIX, the standard that governs Unix time, explicitly ignores leap seconds — which means every Unix-like system has invented its own incompatible workaround.

The Versailles Decision: Who Voted and Who Objected

The 2022 Versailles vote was not unanimous. While the resolution passed with broad support, Russia was the most prominent and vocal opponent of abolishing leap seconds.

Russia’s concern centered on its GLONASS satellite navigation system. Unlike GPS, which incorporates leap seconds into its broadcast offset, GLONASS relies more directly on UTC and would require significant software and firmware updates to function without leap second adjustments. Russian officials estimated it could take up to a decade to fully upgrade GLONASS for a post-leap-second world.

The resolution ultimately passed with the deadline set at 2035 — deliberately long enough for systems like GLONASS to adapt. The United States, France, and Germany were among the strongest proponents of abolition, citing the escalating costs and escalating risks to digital infrastructure that grows more interconnected every year.

Judah Levine, a physicist at the National Institute of Standards and Technology (NIST) in Boulder, Colorado, has been one of the most persistent advocates for ending leap seconds. He designed the NIST time service used across the internet and has argued that the leap second system was designed for a pre-internet era and has not aged well.

The 2035 deadline was a political compromise. Some delegations wanted abolition as early as 2027 or 2030, while others preferred an even later date. All sides ultimately agreed that a firm, binding cutoff was necessary to force action.

The Negative Leap Second: A Crisis Nobody Has Tested

While the world waits for 2035, a more immediate and potentially more dangerous threat looms: the first-ever negative leap second.

Earth’s rotation has been accelerating since roughly 2020, which means UT1 is currently catching up to UTC. If this trend continues, the IERS would need to remove a second from UTC rather than add one. This has never happened since leap seconds were introduced in 1972.

A negative leap second would create a 59-second minute. The timestamp sequence would jump directly from 23:59:58 to 00:00:00, skipping 23:59:59 entirely. Software systems that assume time never goes backward, never repeats, and never skips would face a scenario they were never designed to handle.

Duncan Agnew, a geophysicist at the Scripps Institution of Oceanography, published research indicating that a negative leap second could become necessary within the next few years if current rotational trends hold. He warned that the software world is completely unprepared for this eventuality.

Google’s smear technique cannot handle negative seconds because it was designed exclusively for additions. The entire smearing infrastructure would require fundamental rework. Network time protocol (NTP) servers and operating system time-handling code would need new code paths that have never been exercised in production.

This is why many experts believe the 2035 abolition deadline should be moved earlier — before a negative leap second forces the issue in the most disruptive way possible.

What Happens After 2035: The Future of Timekeeping

After 2035, UTC will no longer be adjusted with leap seconds. Atomic time and astronomical time will gradually diverge, and the world will accept the growing gap as the cost of a stable digital infrastructure.

The divergence rate is roughly one second every one to two years based on historical data. Over a century, UTC could drift about one minute away from true solar time. After several thousand years, the difference could grow large enough that noon would no longer align with the Sun’s highest point in the sky.

The CGPM has tasked a dedicated working group with developing a long-term plan to manage this divergence. Proposed solutions include a “leap minute” — adding 60 seconds at once every few centuries rather than one second every year or two. This would spread the disruption across geological time scales rather than a human lifetime.

Another option is simply allowing the gap to grow indefinitely and addressing it only if practical problems actually arise centuries from now. Given the pace of technological change, future civilizations will likely have very different approaches to timekeeping by then.

For most computer systems, the abolition is purely beneficial. No more emergency patches deployed at midnight, no more weekend monitoring during leap second events, no more vendor incompatibility headaches. NTP servers, data centers, and financial systems will operate on a continuous, unbroken time scale that never skips, repeats, or jumps.

Everyday people will not notice the change at all. Your phone, computer, and every connected device will keep working exactly as they do today, just without the occasional disruptive one-second adjustment that engineers have been quietly managing for over five decades.

FAQs

Conclusion

The end of leap seconds marks one of the most significant shifts in global timekeeping since UTC was created. For over 50 years, these one-second adjustments have kept our clocks aligned with Earth’s rotation, but the cost to digital infrastructure has grown unsustainable.

The leap seconds abolished 2035 decision gives engineers, governments, and system operators a clear deadline to prepare. Companies that build time-sensitive infrastructure should already be auditing their systems and planning for a world where UTC runs continuously without interruption.

If you work in network engineering, software development, or any field that depends on precise time synchronization, now is the time to review how your systems handle time. The negative leap second threat makes preparation urgent — we may not have until 2035 to get this right.