Note sure why you think A will collide with B, or what "collide without contact" means. My interpretation of the scenario is that B moves in the positive $y$ direction, A and C are pulled after it by strings that connect them to B (I am assuming these are light inextensible strings), so throughout the motion A and C have equal velocities in the $y$ direction (until they collide), but also equal and opposite velocities towards each other in the $x$ direction.
I think we must also assume that the size of the balls is small compared to the distance between them, so that just before A and C collide we can treat both strings as being parallel to the $y$ axis. This means that just before A and C collide the $y$ components of the velocities of A, B and C are all equal. However, we need to remember that the velocities of A and C also have equal and opposite $x$ components.
If we call the $y$ component of the velocities of A, B and C just before A and C collide $v_y$ then by conservation of momentum we have
$mv_0 = 3mv_y$
And if we call the magnitude of the $x$ component of the velocities of A and C just before the collision $v_x$ then by conservation of energy
$\displaystyle \frac 1 2 mv_0^2 = \frac 1 2 mv_y^2 + m(v_x^2+v_y^2)$
So you have two equations in two unknowns $v_x$ and $v_y$. Solve these to find $v_x$ and $v_y$, then the velocity of A just before it collides with C is $\sqrt{v_x^2 + v_y^2}$.
Since we are calculating velocities before any of the balls have collided with one another, the question of whether these collisions are perfectly elastic or not is irrelevant.