The major reasons for and against using Drupal 8.x instead of 7.x for a projet due for release soon after 8.0.0 release, as I see them :

Pros:

The lifetime of the 8.*.* versions is likely to be quite long, while actual support for 7.* will decrease like happened for 5.x when 6.x was released, and even more so for 6.x over the years of 7.x. So, to maximize your ability to evolve, stay safe, and find developers at mass market costs, you want to go with the then-current version, that is: 8.0.* when it is released.
Developer profiles are expected to change rather dramatically with 8.*.* : while Drupal was originally dominated by smaller shops and larger ones reluctantly worked it because it was needed but didn't fit their technical practice, Drupal 8 code is difficult to grasp by historical Drupalshops with accumulated lore and no strong software competence. Hence shops succeeding with Drupal 8 are likely to be those existing Drupalshops which invested in updating their knowledge to the level required by Drupal 8, plus the major computing service agencies, for which Drupal 8 becomes just an added RAD and "content modeling" layer on top of a "normal" technical stack instead of an exotic toy. Meaning these are the ones you will want to work with, because they will have ability to execute in a diverse IS landscape instead of boxing your projects within the limits of that single tool.
The mass of new features, especially around headless sites and I18n improvements, makes Drupal 8 overall a compelling proposition.
The new six-months releases policy, à la Ubuntu, means new features twice a year, hence an ability to be made aware and adopt new market needs and trends fast as part of evolutive maintenance.
Cons

A lot of the cost reductions enabled by CMSes over frameworks, and with Drupal specifically, comes from the mass of contributed components (modules, themes). With Drupal 8, the bigest major contributed components have been moved to core (most fields, views...), but the vast majority of the other contributions are not ported yet, and it is likely that many will not yet be ready by 8.0.0. time, meaning projects will have to support the cost of 8.x porting or of developing alternatives. Hence higher costs than a Drupal 7 version are likely at that point in time.
The more technically savvy profiles employed by the up-to-date Drupalshops and major computing service agencies require more initial and more sustained training, so they have higher per-diem rates than the local ad agency turned site builder over earlier versions of Drupal. Hence higher development costs than a Drupal 7 version are likely for the foreseeable future.
The mass of components and original code integrated in Drupal 8 and its relative novelty by 8.0.0 release time mean that updates are likely to be more significant than with 7.x. The new six-months releases policy, à la Ubuntu, means new features twice a year, hence an incitation to apply evolutive maintenance. Hence higher maintenance costs than a Drupal 7 version are likely for the foreseeable future.