The isogeometric scaled boundary approach (SBIGA) combines isogeometric analysis (IGA) with the scaled boundary finite element method (SBFEM). It enables the discretization of 3D solids by partitioning them into three-dimensional NURBS, which are created relative to a scaling center [1]. The scaling center is chosen within the kernel of the domain, from where the entire boundary is visible. This approach can be employed for star-shaped geometries and results in a multi-patch discretization. Based on this approach, the numerical solution is approximated with NURBS both on the boundary and the interior of the domain. The numerical integration is typically performed element-wise with Gauss quadrature. With increasing polynomial order of NURBS, the total number of quadrature points increases significantly. In particular for 3D solids, the computational cost becomes high. Considering the continuity of the NURBS basis across internal element boundaries allows for significant savings. In this work, we explore optimal quadrature rules driven by machine learning to mitigate the computational cost without losing accuracy. Following the idea in [2], the search for a quadrature rule is posed as an optimization problem and solved by machine learning tools. We employ numerical examples to investigate the optimal quadrature rule for isogeometric scaled boundary discretizations. In order to assess its performance, we compare it with element-wise Gauss quadrature in terms of both accuracy and efficiency. [1] M. Chasapi, L. Mester, B. Simeon, S. Klinkel: Isogeometric analysis of 3D solids in boundary representation for problems in nonlinear solid mechanics and structural dynamics, Int. J. Numer. Meth. Engng. (2021), 123: 1228–1252. [2] T. Teijeiro, J. M. Taylor, A. Hashemian, D. Pardo: Machine learning discovery of optimal quadrature rules for isogeometric analysis, Computer Methods in Applied Mechanics and Engineering (2023), 416:116310.