This is an argument, but isn't this where your scenario diverges completely? OpenAI's "means to an end" is further than you state; not initial advancement but the control and profit from AI.

Yes, they intended for control and profit, but it's looking like they can't keep it under control and ultimately its advancements will be available more broadly.

So, the argument goes that despite its intention, OpenAI has been one of the largest drivers of innovation in an emerging technology.