In 1889, delegates meeting in Paris sanctioned a cylinder of platinum and iridium, ninety parts to ten, as the international prototype of the kilogram, nicknamed Le Grand K. Sealed in a vault at the Pavillon de Breteuil outside Paris, its mass equaled one kilogram by definition, not by measurement. The same resolution also sanctioned separate national prototypes, distributed to member countries to calibrate their own mass standards against the original.
A cylinder of metal is a physical object, though, and physical objects change. Even sealed away and rarely handled, Le Grand K and BIPM's own official copies gained mass from airborne contamination and lost tiny amounts each time they were cleaned before a comparison. Because Le Grand K was the definition, its own mass could never be said to have drifted; only everyone else's could. That is what verifications in 1899–1911, 1939–1953 and 1988–1992, plus a special calibration in 2014, found: BIPM's copies diverged from the original by an average of about 50 micrograms over a century, fifty parts per billion. Nobody could say whether the copies had gained weight, the original had lost it, or both, because there was nothing to check either one against except the other.
Fifty parts per billion sounds too small to matter until you trace what stands on it. The kilogram underlies the newton, the pascal and the joule, and every calibration chain running through them. NIST physicist Jon Pratt named pharmaceuticals and electronics as the two industry sectors likely to benefit fastest from that day's package of redefinitions, the kilogram among them.
Metrologists spent decades building a way to stop depending on the object itself. An instrument called the Kibble balance, conceived in 1975 by Bryan Kibble at Britain's National Physical Laboratory, balances a test mass against the electromagnetic force of a current-carrying coil, tying a mass measurement to the Planck constant instead of to any artifact. After laboratories worldwide met a July 2017 deadline to submit their best measurements of that constant, delegates to the 26th General Conference on Weights and Measures met in Versailles on November 16, 2018, and voted unanimously to fix Planck's constant by definition and let the kilogram be derived from it. The change took effect on May 20, 2019, World Metrology Day.
What made Le Grand K worth retiring was not the size of the drift, which by the numbers stayed tiny. It was that nobody could ever prove which side of the comparison had actually moved: the copies were physical artifacts too, so none of them could supply an independent, unchanging standard to check the others against. A kilogram now needs no vault. Any lab with a Kibble balance can derive one directly from the laws of physics, and check its answer against every other lab doing the same.