In the case of developing an exploit, one might reasonably claim that getting 0xdeadbeef in %rip is 90% done. A lot of researchers would stop at that point, because it's sufficient to demonstrate that it's exploitable. Some years ago, before various mitigations, you'd just have to polish up the payload and then you'd be 100%. But today, there might be a fair bit of work turning a demo crasher into a weaponized exploit.

On the other hand, as the immediate goal was to decrypt one specific device and it was regarded as urgent, I would guess that the former is more likely. Your alternative scenario would amount to what would have been possibly the first time ever a software product was completed before it was released.