The OED says that while it is not impossible that the Celtic and Germanic bases are cognate, it is more likely that the Germanic base was borrowed < Celtic in prehistory, alongside the introduction of iron technology to northern Europe.
The OED lists three derivatives for the underlying form:
(i) a derivative of the Indo-European base of ore n.2;
(ii) a derivative of the Indo-European base of ancient Greek ἱερός strong (probably a further sense of ἱερός holy, sacred: see hiero- comb. form);
(iii) a derivative of the Indo-European base of ancient Greek ἔαρ , εἶαρ blood (on account of the colour of iron oxide, especially as seen in iron ore);
But the OED says that these all pose phonological problems and attempts have been made to resolve these problems by suggesting transmission via Illyrian. Borrowing into either Germanic or Celtic from a non-Indo-European language has also been suggested.
The book Language and History in the Early Germanic World (By D. H. Gree) has further discussion on the origin of the word. It is concluded that the Germanic word for 'iron' is unlikely to be native, but rather borrowed from a non-Germanic language.
The closeness of the Celtic isoglosses and the absence of any other counterparts suggests that this source was Celtic, a conclusion borne out by the significant fact of accent doublets in both Celtic and Germanic59, as well as by the linguistic evidence for the dominance of Celtic iron-working over Germania.
59 Birkhan (1970), 141.