The data is held already by nodes.
Only the most recent network data is currently "held" by the nodes, and it is quickly replaced with new data that is coming in all the time,
for the reason that you would otherwise create TeraBytes of data if you were to save/hold/keep track of the ever changing network data of just the past few weeks.
And imagine you do this for multiple altcoins (some of them with high traffic/volume) simultaneously...?
It will be interesting to see what algorithms we (and others) will come up with just to be able to handle the sheer amount of data.
Pruning makes data held by service nodes more valuable.
Only people paid to run nodes will be bothered to keep such large amounts of data, with the added benefits of regional location origins and times of transactions. And when you're getting paid to do something, operating terabytes of data is only factor of profit: if its profitable, you'll do it. If its not, you probably won't.