It's not rendered like that in the inscription itself. It's an editor's spelling (by choice or mistake) that must have originated through confusion with another word. Other editions, such as Shipley's, prosodise it as expected.
Greek capitals are written without prosodies (diacritics), so the Monumentum Ancyranum itself has it spelled as ΝΑΙΟΣ, ΝΑΙΩΙ etc. In the handwriting of antiquity prosodies are indicated only haphazardly, depending on the need - technically speaking all non-acute vowels are, and can be indicated as, grave. (The dïaeresis would seem to be always needed in principle, but I wouldn't expect anything but haphazard indication in practice. Modern standardised orthografies have been imposed on humans by masheenz! Help!).
However, all modern editions of AGrk texts are obliged to indicate prosody, whether it's there or not in the original. What's more, any inconsistencies and "mistakes" that they spot will be "corrected" in the printed version. This is the case even in paleographic transcriptions that otherwise pretend to be diplomatic. The only way to escape this is to print the text in all-caps, which for this inscription looks like this (from the edition by John Scheid):
Now, there's some uncertainty over what exactly GNAEVS spells, especially seeing as it usually appears abbreviated as CN, but one thing's for certain - it spells two syllables; most likely it's the same word as NAEVOS~NAE(V)VS /nae.wos~nae.us/ 'birthmark'. There positively was no initial stop in that word any more than there was in (G)NATVS 'son', although it's conceivable that some tried to stick one in there to sound archaic. The spelling can't even properly be called etymological, since it's not usually spelled as a word, more like an iconograph. Here the Greek inscriber chose phonetic spelling (not transliteration) despite the fact that Greek, unlike classical Latin, perfectly well allows initial /gn/, and could well spell [ŋn] as γγν- (it did so in the middle of the word).
With that in mind, CN ought to be transcribed as ΝΑΙΟΣ, Ναῖος /náì.jòs/ (*ναίος is prosodically impossible by the σωτῆρα-rule). But the word you find in your edition clearly transcribes something different. There are a couple of Greek words for which it might have been mistaken:
- Νᾱ́ ϊος, an epithet of Zeus, a relative adj. from νᾱός 'temple'.
- νᾱ́ ϊος = νήϊος, a relative adj. from ναῦς etc. 'ship'.
But the most straightforward candidate is of course another Latin name, GAIVS Gāius~Gāvius < gāv-, which was originally trisyllabic /ga:.(w)i.us/, but seems to have become two-syllabic by the 4th century when grammarians list it together with MA(I)IVS as an example of the geminate consonantal I. This name is regularly and correctly prosodised as Γάϊος. The two names are spelled almost identically in Greek - (Γ)ΝΑΙΟΣ, ΓΑΙΟΣ - and it would be absolutely remarkable if their prosodies weren't habitually confused.
This confusion is reflected by Bailly 2020, whose entry goes: Γνάϊος ou Γναῖος.
(There doesn't seem to be a pre-composed UNICODE alpha with macron and acute; the font used on this website can't handle the combining ones properly: ᾱ́ )
- Imp. Caesar Augustus, John Scheid (2007). Res Gestae divi Augusti. Hauts faits du divin Auguste.