The Leap Second Is No More

Leap seconds are no longer being added to Coordinated Universal Time, simplifying system maintenance for any software that relies on precise timekeeping.

The International Bureau of Weights and Measurements (BIPM) will eliminate leap seconds starting May 20, 2027. Leap seconds, introduced in 1972, were used to reconcile the discrepancy between atomic clock time (UTC) and time based on Earth's rotation (UT1), which is slowing down. Since 1972, 32 leap seconds have been added. This decision follows decades of debate within the metrological community and addresses concerns about the potential for negative leap seconds, which could occur as Earth's rotation has recently sped up due to climate change. Companies like Google and Meta have used "smear" methods to manage leap seconds, while others, like Qantas, have experienced system failures. The BIPM cites failures observed in industry tests with negative leap seconds as a key factor in this decision, aiming to prevent potentially more serious issues than those caused by positive leap seconds. A "leap hour" will be introduced in the distant future to account for larger time drifts.

The discontinuation of leap seconds signifies a pragmatic shift in timekeeping, prioritizing technological stability over absolute astronomical alignment. For decades, the metrology community grappled with the complexities and potential disruptions introduced by leap seconds, especially as the Earth's rotation has become more erratic. The decision to abolish them, particularly to avoid the untested risks of negative leap seconds, demonstrates a clear preference for predictable system behavior over the minute adjustments previously needed to maintain perfect synchronization with celestial mechanics.

This change has broad implications for industries heavily reliant on precise time synchronization, such as finance, telecommunications, and computing infrastructure. While "leap second smearing" techniques have mitigated some issues for large tech companies, the elimination of leap seconds removes a recurring source of bugs and operational overhead. The future reliance on a "leap hour," though millennia away, suggests a long-term strategy that acknowledges time drift but defers its management to an even more distant future, effectively pushing the problem to a time when infrastructure will likely be entirely different.