The extraction of the QCD coupling $\alpha_s$ from the comparison of experimental data on inclusive W and Z bosons hadronic decays to state-of-the-art perturbative QCD calculations %at N$^{3}$LO accuracy
is reviewed. The relatively small amount of W data from $e^+e^-\to$\,W$^+$\,W$^-$ collisions at LEP leads today to an non-competitive extraction of the strong coupling at the Z mass from the measured $R_W$ ratio of hadronic-to-leptonic branching fractions,
$\alpha_s(m_Z) = 0.117 \pm 0.042_{exp} \pm 0.004_{th} \pm 0.001_{par}$ with a $\sim$35\% propagated experimental uncertainty. Analysis of the much more abundant hadronic results at the Z pole leads to $\alpha_s(m_Z) = 0.1203 \pm 0.0030$, with a 2.5\% uncertainty by combining three different pseudo-observables (ratio of hadronic-to-leptonic widths $R_Z$, hadronic peak cross section
$\sigma^o_Z$, and total width $G_Z$). An $\alpha_s$ determination with per mille uncertainty requires high-statistics W and Z bosons data samples at future $e^+e-$ colliders, such as the FCC-ee, combined with even higher precision (N$^4$LO) pQCD calculations.