What are the SI units actually based on now?
Since a redefinition that took effect in 2019, all seven base units are defined by fixing the numerical values of physical constants — not by physical objects, and not by properties of materials. It is the most significant change in measurement since the metric system was created.
The problem it solved. The kilogram was, until 2019, defined by a physical cylinder of platinum-iridium held near Paris. Its mass was the kilogram by definition — so if it changed, the unit changed with it. And it did: comparisons with official copies showed divergence of tens of micrograms over a century, and nobody could say which had drifted. The last artefact-based unit was an embarrassment, because the standard could be dropped, contaminated or lost.
How the units are defined now:
The second — by the frequency of a specific caesium atomic transition.
The metre — by fixing the speed of light at an exact value, making the metre the distance light travels in a defined fraction of a second.
The kilogram — by fixing the Planck constant, realised in practice with a Kibble balance that relates mass to electrical quantities.
The ampere — by fixing the elementary charge.
The kelvin — by fixing the Boltzmann constant, replacing a definition based on the triple point of water, which depended on the isotopic composition of the water used.
The mole — by fixing the Avogadro constant as an exact number.
The candela — by fixing a luminous efficacy value.
Why this is better:
The definitions cannot drift, because constants do not.
They are reproducible anywhere, by anyone with the right apparatus, including in principle off Earth — no pilgrimage to a vault required.
Precision can improve indefinitely as measurement technology advances, without redefining anything.
What did not change: the size of any unit. A kilogram remains a kilogram — the definitions were chosen to be continuous with the old ones, so no measurement anywhere became wrong overnight.