Why is anything "necessary"? Who gets to define that? The biological imperative, if you will, is to survive, reproduce, and exploit every niche. Look all over the planet, and you will see that living systems have done exactly that, to a degree well beyond human engineering. If space is a new niche for humans, especially other planets, then why shouldn't humans populate and exploit that environment? From this view, you could say it is as necessary as extremophiles living on undersea thermal vents or inside rocks hundreds of meters below the surface, at a metabolic rate so low it may take thousands of years for them to reproduce just a few times. Nobody has written up a TODO list that demands life fill these niches, and yet, here we are...
We take it for granted that planet Earth will be available as our home for as long as we care to think about the prospect. However, the way we are living right now makes it clear that this is an unreasonably optimistic presumption. Making space habitable seems like a pretty smart way to hedge any bets about our future as a species.
Also, if it turns out that we are not alone in the galaxy, and a hostile alien race visits our planet, the odds are that we will not be able to defend ourselves. At that point, our only hope will be that enough of us escape to the stars to rebuild elsewhere, or seek help.
Suppose we develop a super-intelligent AI which decides for itself to expand beyond Earth and into the stars. Surely such a being will build robots and do things the smart way, right? Perhaps. But why doesn't such a being exist already, when we have petaflops of computing power available to us? One reason is that while our robotic (and computing) technology is incredibly advanced, our biological technology is even more advanced. Those petascale machines operate with electrical budgets measured in megawatts. And they offer roughly the same scale of raw computing power as your little 3 pound brain, which sips a mere 20 W of energy. The next time someone calls you a "dim bulb", you should say: "I sure am! But look at all this dim bulb can do!"
When Curiosity or Spirit or Opportunity suffer a malfunction or failure, scientists just try to make do with whatever systems are left working. The name of the game is as much redundancy as we can afford, and limited expectations for usable service life. The official mission duration for Curiosity was 2 earth years. If humans could only offer 2 years of useful working life, we would consider that an utter failure.
If a super-intelligent AI wants to travel the stars, why would it not use the best technology available? That technology is not offered by Boston Dynamics, as impressive as their offerings have become. The only truly adaptable, self-healing, energy-efficient, high strength-to-weight ratio nanotechnology exploration machines we have available today are humans. That super-AI will immediately recognize that DNA-based life is the pinnacle of nanotechnology and energy efficiency, and will build its endeavours around that technology. Humans may only be the starting point for what such an AI would send out into the stars, but I find it to be an infinitely more likely starting point than mere robots.