We call S(u)$S(u)$ the space complexity of the vEB tree holding elements in the range 0$0$ to u-1$u-1$, and suppose without loss of generality that u$u$ is of the form 22k$2^{2^k}$.
It's easy to get the recurrence S(u2) = (1+u) S(u) + Θ(u)$S(u^2) = (1+u) S(u) + \Theta(u)$. (In Wikipedia's articlearticle the last term is O(1)$O(1)$, but it's wrong because we must count the space for the array.)
Van Emde Boas (and others) gave in [1] the trivial bound S(u) = O(u log log u)$S(u) = O(u \log \log u)$, and later in [2] he found a clever way to combine the data structure with otheranother one so thatin order to reach space complexity O(u)$O(u)$, while mantainingmaintaining the O(log log u)$O(\log \log u)$ time bounds.
But, modern references present the original data structure and state without proof that the space complexity is O(u)$O(u)$. For instance, the very recent 3rd edition of "Introduction to algorithms" by Cormen et al. (ZBL1187.68679) leaves it as an exercise.
I tried with some friends to [dis]prove the O(u)$O(u)$ bound without luck.
[1] http://www.springerlink.com/content/h63507n460256241/
[2] http://www.cs.ust.hk/mjg_lib/Library/VAN77.PDF
van Emde Boas, P.; Kaas, R.; Zijlstra, E., Design and implementation of an efficient priority queue, Math. Syst. Theory 10(1976), 99–127 (1977). ZBL0363.60104.
van Emde Boas, P., Preserving order in a forest in less than logarithmic time and linear space, Inf. Process. Lett. 6, 80–82 (1977). ZBL0364.68053.