Couple of other things. Elsevier is $3 billion in revenue, not $10 billion (the author must be thinking of Reed Elsevier here, Elsevier's parent company.) So really they only have something like $3 billion reasons to not let the internet drive the costs of scientific publishing down to zero.

And to be clear, so far as I can tell, SOPA has almost nothing to do with arxiv.org, because arxiv.org publishes the manuscript, not the final version of the article. Elsevier's stated position is that publishing the manuscript version is not copyright infringement.

Even if SOPA passed, sure, it would be easier for Elsevier to shut down arxiv.org. But if arxiv.org started publishing the final version of articles for which Elsevier had obtained worldwide exclusive copyright, that would be copyright infringement anyway. So the key (it would seem to me) is whether arxiv.org publishes the final version of articles or not (they curenttly don't), not whether SOPA passes, that determines whether arxiv.org is to get shutdown.

And so if they did, the answer to the question "Could Elsevier shut down arxiv.org?" would be "Yes."

Right now, the answer is "No."