"Nothing exists" is one (tongue in cheek) way of thinking about the axiom of the empty set.
The empty set $\emptyset$ exists by fiat; which is to say (either directly as an axiom or derived from other axioms), $\exists \emptyset : \forall x, x \notin \emptyset$.
Once we have any set $x$, we have rules by which we can prove the existence of other sets; such as the set $\{x\}$ whose only element is $x$, and so on. $\emptyset$ is no different in this respect; it is a set, and we have rules that show how to construct other sets from sets whose existence we have already proven.
The union of two sets $A$ and $B$ is a set with the property $\forall x: (x \in A \cup B \iff x \in A \lor x \in B)$. So thinking about $\emptyset \cup A$, we are talking about a set:
$$\forall x: (x \in \emptyset \cup A \iff x \in \emptyset \lor x \in A)$$
But by definition, it can never be true that $x \in \emptyset$; so the above boils down to:
$$\forall x: (x \in \emptyset \cup A \iff x \in A)$$
or in other words (see axiom of extensionality); $\emptyset \cup A = A$.
On the other hand there does exist some set $x$ s.t. $x \in \{\emptyset\}$; namely, the set $x = \emptyset$. So when we consider $\{\emptyset\} \cup A$,we are talking about the set:
$$\forall x: (x \in \{\emptyset\} \cup A \iff x \in \{\emptyset\} \lor x \in A)$$
Thus $\{\emptyset\} \cup A$ is not necessarily equal to $A$; because (depending on $A$) we might not have $\emptyset \in A$, but we do have $\emptyset \in \{\emptyset\}$; and therefore $\emptyset \in \{\emptyset\} \cup A$.