This is definitely not linear, and keep it mind it's only with a small amount of early adopter that they reached 125 Go mark.
As a matter of comparaison the whole database of the earth map at openstreetmap.org is as of writing only 36 Go fat when delivered in compressed form. https://planet.openstreetmap.org/pbf/
It seems absolutely ridiculous and a waste of ressource to expect that the blockchain shall never/ever use some rotating process to loose weight at some point.
Blockchain like technology can be very usefull, but current implementation look like a very rotten seed. And if a newer version that solve those issue emerge why on earth would they want to give any value to a prototype currency driven by opportunistics "first in" miners?
I desperately hope that no public fund was stupid enough to invest any real life saving to this mess because IMHO it's not a matter of if, but a matter of when it will collapse.
Now to be more accurate it might rather be a growth according to a logistic curve. Right now Bitcoin is used only by early adopters, but should Bitcoin usage diffuse among general customers it will growth quasi exponentially before plateauing.
A hundred years from now, you could store the blockchain on current day hardware. And that is without talking about improvements in technology.
I mean if Bitcoin ever gain real traction isn't it pretty inefficient to store all the data and replicate it around? Storage now seem infinite but practicality it is not, won't current blockchain be saturated to the point of uselessness when/if it reach usage level of an actual banking network?
"Let think about scaling later" is maybe not a good idea for a currency implementation...