The numbers are "what's applied to the device". In switching, you spend most of your time at e.g. 300V ~0A, or 5A ~3V (in Rds(on) mode), with only a few short instants at e.g. 300V 5A, or ~3V ~0A.
You're also looking at the 10ms curve, which isn't DC (this device is not rated for linear DC operation -- don't use it in a HV current sink!). The 10us curve would likely be more representative, but even then, since only the switching edges are dissipating much power, an even shorter time curve would be relevant, like 100ns or less! At these time scales, heat doesn't even leave the junction, and the die thermal mass can be treated strictly as a capacitor (see also the transient thermal graph for t < 20us). So, no worries.
Tim