As there has been no answer so far, I would say that it is not attested.
I have never encountered it in texts or grammars — and I would be glad to hear whether more experienced Latinists share my experience.
It seems consistent in Latin that a verb can be treated as a noun, which uses the infinitive for nominative and accusative (without prepositions) and forms of gerund for the rest (including accusative with prepositions).
As the infinitive seems to be usable in all uses of a nominative, there seems to be no use for a nominative gerund.
Of course, this does not rule out the nominative gerund being used instead of the infinitive, but I also see no compelling reason to ever do so.
(I would treat the combination of the infinitive and the gerund a single noun on a semantic level, but not on a morphological one. This seems to be a matter of taste.)
The example Alex B. mentioned was constructions of the form mihi scribendum est.
The question is whether scribendum should be analyzed as a gerund or a gerundive.
When a noun is used, as in liber mihi scribendus est, all (I think) will agree that we have a gerundive instead of a gerund.
The canonical interpretation seems to be that even without a noun it is still a gerundive.
But given how the lack of a noun can turn a gerundive into a gerund in some cases, I can see how some would read it as a gerund instead.
When it comes to questions like this, I am not sure whether one should make a distinction between gerunds and gerundives in the first place.
One could simply take gerundive as a verbal adjective and the gerund as substantivized neuter version.
Such substantivization of adjectives is not rare.
If we take this interpretation of Latin gerunds, the question whether scribendum is gerund or gerundive seems to be whether the nominative of a substantivized adjective (whose attestation is unclear) is truly different from the neuter nominative of the adjective (which is amply attested).
I am not sure whether the question is even meaningful; it seems to be a matter of opinion, as I see no functional implications.
For the distinction and origins of gerunds and gerundives, see this question.
Even if you keep a distinction between gerunds and gerundives, it is not clear what it means to decide that the scribendum in scribendum est is a gerund instead of a gerundive.
To me it seems sensible to interpret Latin grammar either way, as there seem to be no visible implications in use nor major contradictions in theory.
Yes, the nominative gerund is attested if you decide that the scribendum in scribendum est (without a noun) is a gerund.
Otherwise, I am not aware of clear cases like *volandum difficile est, and I think they are not attested.