We show that a broadly accepted criterion of laser-induced breakdown in solids, defining the laser-breakdown threshold in terms of the laser fluence or laser intensity needed to generate a certain fraction of the critical electron density rc within the laser pulse, fails in the case of high-intensity few-cycle laser pulses. Such laser pulses can give rise to subcycle oscillations of electron density $h̊o$ with peak $o̊$ values well above $r$̊c even when the total energy of the laser pulse is too low to induce a laser damage of material. The central idea of our approach is that, instead of the $rh̊þinspace=þinspace$ρr̊atio, the laser-breakdown threshold connects to the total laser energy coupled to the electron subsystem and subsequently transferred to the crystal lattice. With this approach, as we show in this work, predictions of the physical model start to converge to the available experimental data.