The usual theory I've heard isn't so much that fancy starchitect headquarters
cause the companies to fail. Rather, it's more of a signaling story. To wit: new headquarters say something about the kind of company you have, and
what it says is usually a bad sign.
When companies are doing really well the corporate offices are too busy doing stuff to hire a starchitect and the company is growing too fast for it to make sense to literally set in stone how it should henceforth be arranged.
You tend to build a fancy new building when the company has stopped growing (so you know how big to make the building and how to lay out the various departments) and when the management staff is to some extent twiddling their thumbs looking for something useful to do rather than starting bold new projects. That sort of building signals the company is coasting rather than being dynamic. They are investing time and money and effort in a building because (unlike when they were growing) they currently can't think of anything better to invest that time and money in.
However, both Apple and Google are arguably exceptions to this rule.
In Apple's case, mitigating circumstances include that they've
already survived building and moving into the "Infinite Loop" campus and a vast flood of architecturally-interesting Apple Stores. Their new campus - while impressive - is essentially iterative; it's doing something they've done before using skills they've successfully used before, merely on a larger scale.
In Google's case, their (goofy) design explicitly seems to anticipate LOTS of change and potential further growth and reorganization. They are not "setting in stone" the shape of the company.
And in both Google AND Apple's case, the companies have so much money and so many employees that it'd be stupid NOT to try to consolidate a bit at this point.