The Szász inequality is a classical result that provides a bound for polynomials with zeros in the upper half of the complex plane, expressed in terms of their low-order coefficients. Generalizations of this result to polynomials in several variables have been obtained by Borcea-Brändén and Knese. In this article, we discuss the Szász inequality in the context of polynomials with matrix coefficients or matrix variables. In the latter case, the estimation provided by the Szász-type inequality can be sharper than that offered by the von Neumann inequality. As a byproduct, we improve the scalar Szász inequality by relaxing the assumption regarding the location of zeros. Finally, we estimate the Agler norm of a scalar multivariate polynomial.