i don't know about this 100w per plant thing... 21-22 DLI is sthe minimm you want, and area of coverage is relative. it can amount to less than 100w in many contexts.
single point light, maybe 50w? if efficient and you kept plant fairly small footprint, it might do okay.
efficacy is often lied about on specification sheets, unfortunately. a 2umol/J light needs 150% more watts to provide same amount of light as a 3umol/J light. And that's a very possible context, give or take a few decimals. Watts translates to different outcomes due to differing efficacy. watts are a horrible thing to rely on... when i refer to watts it's simply because it's a slightlay more understood concept that umol/s, umol/J, ppfd, or DLI... all of which can paint a much more accurate picture.
read up on DLI. it'll allow you to compare apples to apples. most the other terms should be covered too. you want 21-40 DLI range.. higher the better.. there is a ceiling at some point based on environment and genetics to a degree too. lower range will have higher proportion of larf... higher range will have a better yield and lower proportion of larf. if you provide much less than 21dli, the plant will not be able to add mass... it may even contract or die.
in future when buying a light...
work backward from area of coverage... how many sq ft? if metric, even better you won't have to convert anything. reference a DLI table - google image search or make one they are simple as pie. with atmospheric co2, your ceiling is somewhere near 40 DLI. you don't "have" to max out, of course.
If using autoflowers, you can use 16-24 hours of light... if using photoperiod plants, you need to focus on teh "12-hour" use row. Find "40 DLI" give or take in those respecive rows. this will give "PPFD" which is just umol/s per m^2. Factor for your size of garden... if smaller than a 1m^2 area, you will need less than the value in the DLI as umol/s produced by a light.
so if .5m^2, you need 1/2 of the listed PPFD in umol/s produced from the light... let's say oyu have photperiods and 12/12 cycle... that's roughyl 900PPFD... you'd want 450umol/s from your light. the height from canopy should be based on best coverage of entire area.. observe and react, this is only a good starting point.
5-10% extra is never bad. LED like any other light dims over time. higher quality LED diodes should take 50k hours to dim to 90% of original intensity. this gets a bit complicated as spec sheets lie their asses off, again... you often have to reference the "real" spec sheet at samsung.com or bridgelux etc... their spec sheet is accurate, but it is also 1 diode type out of many in any 1 model, and it is the "best" one, of course but not the best for a garden... warmer white light and other factors reduce this optimal number given by the original manufacturer. but, if you get into nitty gritty, good to look at these sheets and determine how hard a particular diode "should" be driven to hit advertised specifications.
e.g. an LM301b or h samsung diode that isn't driven near 0.25watts per diode (simply divid watts/diode count) to be anywhere near 2.7-3.0 umol/J efficacy. watts multiplied by efficacy equals umol/s, if 1 of these specs is not advertised, you can calculate them easily. if it deviates greeatly from samsung's sheet, you know they lied their fuckin asses off.
cheap lights lie 99% of the time, fyi... even "reputable" grow light brand names in mid range often lie their asses off too. it's good to learn what the diodes are capable of. then you can weed through the bullshit.