I think one suited approach to your problem is to look at the stability of the movement with respect to the energy.
Energy stability analysis
The rotation energy is given by
$$E_{rot} = \frac{1}{2}\Theta{(\mathbf{\omega},\mathbf{\omega})}$$
where $\Theta$ is the (tensorial) moment of inertia with respect to the (vectorial) rotation $\mathbf{\omega}$.
The mass of a prolate, homogeneous ellipsoid is given by $M = \frac{4}{3}\rho r_a^2r_b$ where $\rho$ is the mass density and $r_b$ is assumed to be the axis longer (or, say different) than $r_b$.
Then, we can write down the principal moments of inertia for this ellipsoid as
$$\Theta_b = \frac{M}{5}\cdot 2 r_a^2$$
and
$$\Theta_a = \frac{M}{5}\cdot (r_a^2 + r_b^2)$$
If we now chose $\theta$ to be the angle between $\mathbf{\omega}$ and the $z$-axis, we find the rotation energy as
$$E_{rot} = \frac{M}{10}\left( 2r_a^2 \cos^2\theta \omega^2 + (r_a^2 + r_b^2) \sin^2\theta \omega^2 \right)$$
with $\omega$ now being the scalar value of rotation (I could not make it bold).
Now, we have to turn to the potential energy in the gravitational field. I suppose, this should take the form
$$E_{pot} = Mg\left( r_b \cos\theta + r_a \sin\theta \right)$$
If you now analyze the energy
$$E = E_{rot}+ E_{pot}$$
with respect to $\theta$ and the parameters $r_a$ and $r_b$, you will find out that $\partial_\theta E$ will vanish at $\theta = 0$ but it will be a maximum (remember, $\theta\geq 0$) assuming $r_b > r_a$.
That means, that the movement will be unstable assuming a rotating prolate body. Or, in other words, you cannot find some $\omega_s$ to have the egg at resting rotation in this model.
Other approaches
One further ansatz to calculate a stable rotating egg, one could assume that the rotating egg is not homogeneous. A simple model could be to assume that a small ring of mass is attached to the "smaller belt" leading to
$$\Theta_b = \frac{M}{5}\cdot 2 r_a^2 + \frac{M_{ring}}{2}r_a^2\quad, \qquad \Theta_a = \frac{M}{5}\cdot (r_a^2 + r_b^2) + \frac{M_{ring}}{2\cdot 2\pi r_a}r_a^2$$
Now, can you show, when $E$ has a minimum in $\theta = 0$ leading to a stability condition of the form $F(M_{ring}, r_a/r_b; \omega_s) = 0$?