It’s annoying. That’s what it is. A tiny, invisible hiccup in time that brings down servers.
Global timekeepers are getting ready to vote. The proposal? Kill the leap second. Replace it with something far bigger: a leap hour. We might see this as early as 2027, which sounds ridiculous until you consider what the alternative costs.
The Broken Clock
We’ve been doing this since 1972. The International Bureau of Weights and Measures (BIPM) invented the leap second because Earth’s rotation slows down, gradually, unpredictably. Coordinated Universal Time (UTC) stays rigid, atomic, perfect. The Earth wobbles. So once a year, maybe less, we add one extra second to catch up.
It works on paper. On computers? It’s a disaster.
“Always been a problem.”
That’s Judah Levine, former physicist at the US National Institute of Standards and Technology, summing up forty years of temporal housekeeping.
Modern code hates sudden stops. Messages arrive out of order. Logs get corrupted. We’ve seen Meta crash. Reddit go dark. Cloudflare stumble. Planes grounded. Stock traders sweating over milliseconds lost. As networks demand nanosecond precision, the leap second is a stone in the shoe. Developers stopped listening to the BIPM years ago, implementing their own patches, smear seconds that dilute the extra time across hours rather than slamming it all in at once.
The Negative Threat
Here is where it gets worse. Since 2016, Earth stopped slowing down. It started speeding up.
No leap seconds added. Just silence.
But now we face a different problem. If Earth spins faster, UTC falls behind. To fix it, we would need a negative leap second. Delete one second from existence.
Has that ever been tested? No. Will it break things? Probably everything. We don’t know, but everyone assumes catastrophic failure. The risk isn’t hypothetical; it’s looming.
The Hour Fix
Timekeepers have been dragging their feet, holding conferences every four years, debating thresholds. In 2022, they agreed to loosen the rules by 2035, allowing UTC to drift further from Earth’s spin without correction. A safety net.
Projections say that won’t matter. The clock is ticking. Before 2035, the math says we might need that dangerous negative leap second.
Enter the leap hour.
Instead of fixing time every year with a single, jarring second, we add—or subtract—an entire hour. Once. Maybe in two centuries. It’s a blunt instrument, yes. But it’s stable. It lets UTC and Earth rotate without constant interference, smoothing out the jagged edge that crashes software.
Patrizia Tavella, director of Time at BIPM, isn’t interested in waiting until 2035 to be safe.
“If we wait till 20235, we have 330 percent risk.”
Thirty percent chance of a disaster? Too high. Tavella asked stakeholders what they thought. What if the risk dropped to ten percent? They still said no. Not enough safety margin for the backbone of the digital world.
The goal is 2027 implementation. The vote approaches. We’re about to stretch an hour across the globe. Or compress it.
