Look at the pulsed current rating of the IRF2804--1080 amps!
Or how about the power dissipation rating--330 watts in a TO-220 package? Try that (continuously) without a water cooled heat sink (or heat pipe, or something equally able to remove large amounts of heat without a large temperature rise)!
Once upon a time, I wondered just what current it would take to melt off the leads of a TO-220. I took a TO-220 device that had been blown, and cracked the epoxy off and scraped off the leads to the silicon. I took a heavy copper clamp and clamped it to the tab of the package. I put another clamp out on the very tip of the center lead, the one that is galvanically connected to the tab. I then applied a large current and increased it until the lead melted. It melted at about 150 amps, and the melting occurred in the corners where the lead necks down (current crowding, no doubt). If you make a connection closer to the package, before the decrease in cross sectional area of the lead, and remove heat there, you *might* be able to do 150 amps continuously. Good practice would dictate accepting the manufacturer's recommendation of 75 amps, package limited.