Most answers/comments seem to have the philosophy that "there is nothing to lose by submitting to the top journals, so go for it". I would like to offer a different perspective.
There are costs to every journal submission.
Time spent formatting the manuscript to fit the journal. This sometimes requires relatively minor changes such as formatting references, but might involve more significant effort such as adapting the length to fit a word limit, or altering the focus of the introduction. A lot of the very top journals have quite specific style requirements that are very different to other journals.
Time waiting for a response. At least if it is a desk reject, this will be quick. If it's a rejection after reviews, then as Bartosz points out, at least you will get some good feedback. But the wait can be important at certain stages in a scientific career. One paper can make a big difference to a grad student's CV.
Potential (but often very real) emotional cost of (possibly repeated) rejections.
Costs to the system. If everyone adopts a policy of always submitting to the highest journals, editors become swamped and even more likely just to reject on sight. (Yes, I know, it will still happen because of the Tragedy of the Commons).
These costs need to weighed against the potential benefits if the paper is accepted, which of course are large, but far from guaranteed in the case of high-ranking journals.
Therefore, I don't think it is as simple as saying "you might as well try, you've got nothing to lose". There comes a point where the chance of acceptance drops so low that it is not worth sinking the above costs into an effort that is very unlikely to succeed.
Notice that the costs listed above are much reduced or absent for the supervisor. Therefore, while there is little harm for the supervisor in aiming high to start with, the same is not necessarily true for the student. Therefore, it may well be worth getting some further opinions as to the chance of success for the submission. Ideally, try asking some other faculty members whom you know and trust to have a quick read of the manuscript and give their opinion. They may agree with the supervisor, in which case, great, go ahead and try! But it is worth asking around before putting in all that effort.
NOTE: This is not meant to sound negative. The answer for your friend may well be that he should give it a go. But I do think that there is a non-trivial decision to be made.