the invariance of the speed of light
This is the strange, revolutionary half of Einstein's postulates. It says that light in empty space always moves at the same speed — c, about 3.00 x 10^8 m/s (300 million metres per second) — for every inertial observer, regardless of how fast the source of the light is moving or how fast the observer is moving. Everyday intuition screams that this cannot be right: if you chase a departing car, its speed relative to you drops. Chase a light beam, though, and it still recedes at the full c.
Precisely: the speed of light in vacuum is invariant, meaning it has the same numerical value in every inertial frame. Ordinary velocities add the way common sense expects (a ball thrown forward at 5 m/s on a train doing 20 m/s moves at 25 m/s over the ground). Light stubbornly refuses to play along. Whether the flashlight is racing toward you or away, whether you sprint toward the beam or flee it, you always measure the light approaching at exactly c. This isn't a limitation of our instruments; it is how spacetime is built.
The consequences are enormous. If everyone must agree on light's speed — distance divided by time — but observers move differently, then they cannot all agree on distances and times themselves. Something has to give, and what gives is the absolute, universal clock and ruler we assumed since Newton. Time dilation, length contraction, and the relativity of simultaneity are all the price of keeping c constant.
Light reaches us from a distant star and from a galaxy rushing away at a large fraction of c. Measured here on Earth, both arrive at exactly the same speed c. Their motion changes the light's colour (a Doppler shift), but never its speed.
Source motion shifts light's colour but not its speed — c is the same for everyone.
The constant c is light's speed in vacuum. In glass or water light travels slower, and objects can even move faster than light does in that medium (which is what causes the blue Cherenkov glow) — but nothing ever beats c, light's vacuum speed.