SPF50 means that only 1⁄50 of total UV light gets to your skin. If the SPF50 factor for a narrow light wavelength was tested under 20 μm layer of sunscreen, and you’ll manage to apply 40 μm layer, the resulting UV light would be 100%÷50÷50=0.04%, corresponding to SPF2500 for this wavelength.

Of course, SPF is not measured for one particular wavelength. UV light is a spectrum of sunlight of wavelengths starting at 100 nm to 400 nm. UV radiation of shorter wavelengths contains more energy, but is easier to stop; it usually responsible for short-term damage like sunburns. For longer waves it’s vice versa, they are less powerful and won’t cause a burn, but are harder to stop and penetrate skin deeper and cause cancer and sun damage.

Low-SPF sunscreens are usually achieving their index thanks to stopping easy-to-catch short waves that are most of UV exposure in pure joules, and are having a hard time stopping long-wave UV due to their chemical composition. For further reading I recommend a 2019 article “Critical Wavelength and Broad-Spectrum UV Protection”, Google it, I’m cautious to paste a link in case anti-spam bots are tightly configured here.

SPF is overall not a very good and all-encompassing sunscreen quality indicator. I recommend looking into UVA-centered indexes: CW (over 370 is good, in EU these are marked with UVA circled mark), UVAPF or PPD (over 15 is decent, that’s same system as SPF but for UVA, longer waves) or, common on asian brands, PA++… index (UVAPF=2^x, where x is the amount of “+” symbols after PA).