Tough problem. Finding numbers like that is actually a problem about elliptic curves, which have a lot of deep mathematics surrounding them. These are elliptic curves because you can write them in the form y^2=x^3+ax+b:
y^2=x^3+2
y^2=x^3-2
(the second one is what applies in this example but the first one also fits your description.)
So an approach using the theory of elliptic curves might help. One thing that appears to help from the Wikipedia article is a theorem that says that there are only finitely many solutions.
I did a straightforward numerical method, just setting a number equal to sqrt(x^3+2) and putting out x and that square root if the result was an integer, and amazingly there are no solutions other than this one for x<100,000, but a lot for 100,000 < x < 1,000,000: 517 (counting the first one), in fact. Going up to 10,000,000 there are 2696 solutions (counting the first one). So 26 is pretty special until numbers get massive (x>100,000 means that the first number of interest is in the quadrillions (!)).
The other one (where the cube is the smaller number) has no solutions for x<100,000 but has a pretty fair number once we get above that (comparable to the previous one actually).