http://referenceworks.brillonline.co…_COM_vol2_0005
Should the stress fall on the third syllabus?
The “ī” is long (because of energīa from ἐνέργεια ), so it must carry the accent.
Check my recording: http://linguax.com/energiicus_godmy.mp3
(if the second to last syllable is long either because of two consonants following the vowel in that syllable or because of a long vowel, it must carry the accent, otherwise the third one carries it)
Godmy is correct in stating rule about Latin pronunciation. However, I cannot find energicus in any Latin dictionary.
It is a new derivation, no doubt (but it doesn’t stand in our way in solving this problem (or in understanding it) 
)
The trouble is the i is not necessarily long in energicus just because the i of energia is long.
What happens with magia/magicus?
For what is worth, in Italian we say enérgico (and mágico)
You are right, on the other hand: the derivation of the magicus adjective from magīa was done here already in Greek in μαγικός . With energīcus/energicus we derive it purely in a Latin environment (though you could theoretically think of the derivation here happening through Greek too, analogically), so I wouldn’t rule out the ī completely… (it depends if you think of the Latin suffix just as about “-icus” or “-cus”)
Anyway, here is an alternative recording with a short “i” and an accent on the third to last syllable (accordingly): http://linguax.com/energicus_godmy.mp3
If the -icus of energicus is modelled on the Greek adjectival termination ικός (and clearly it is), then the only sensible choice is for the i to be short. Gildersleeve, among others, clearly indicates that this is the case:
http://books.google.co.uk/books?id=Qk7foW63cO0C&pg=PA225&dq=latin+energicus&hl=en&sa=X&ei=VKQEVMGiEoOL0AWlq4HwCQ&ved=0CCQQ6AEwAA#v=onepage&q=latin%20energicus&f=false
Alright. I have posted a modified recording already. The OP can choose, listen and read the arguments in the same time…