Heat will always be a factor when you are using power amplifying transistors - just look at the heatsinks and ducting fans inside any commercial PA audio amplifier like a Peavey, QSC or Crown for TO4 case transistors.
Now, the transistors in the PMD are indeed, 500 Watt transistors, in a push-pull arrangement. Now, that being said, despite comments on here elsewhere and in advertisements for certain PMD cooler vendors, they are not putting out 500 Watts of power/heat each! Not even close to it, by far! If they were, not only would they literally melt the circuit boards inside the PMD, they'd boil the diesel fuel inside the IP before it would make it through the pump. That's a 1,000 Watts total, the power of one of your burners on an electric stove!
They are switched on/off and are rated at 500W so that it is a clean power signal that does not clip. Look at the gauge of the wires in/out of the PMD plug. The continuous amperage draw and send from those two transistors couldn't be handled by those wires. The actual continual draw is about 30 Watts of the PMD, the total heat output of the two transistors is the same.
The two biggest problems with the OEM Stanadyne PMD are a result of GM's pencil pushers. Stanadyne designed their driver (the PMD) to be remotely mounted away from the IP and with a sink/radiator assembly (sound familiar?) The pencil pushers, accountants and engineers, the former balked at the additional costs associated with the parts required, the latter came up with the "solution" - over the protests of Stanadyne's engineers - of mounting the PMD to the side of the Stanadyne IP with the idea of satisfying the bean-counters by eliminating wiring and a heat sink while using the mass of the IP and the "cool" fuel flowing through it as a heat sink and dissipation method for the heat produced by the two power transistors. Only two problems with that. The IP is located in arguably one of the hottest locations in the engine compartment, in the valley with heat sources all around it (coolant crossover, intake manifold, heads) and then already poor airflow to cool it made even worse by that stupid plastic 6.5 Turbo engine cover.
Stanadyne's pencil pushers were also at work as far as the PMD was concerned, using standard tolerance components and only a basic design in the first series of "cigarette pack" black ones. Later versions of that one made by Stanadyne came out in the early 2000's added things like a limp mode that didn't shut the injection signal off completely and kill the engine when the PMD overheated, but that was mainly because GM/Stanadyne was tired of replacing IPs under warranty from DTCs for "failed" IPs when the PMD fried and vehicles were dying in to middle of nowhere or the middle of a traffic jam.
Late 2000's/early 2010's Flight Services, upon the expiration of Stanadyne's Patent, began addressing the problems of the OEM unit's failures, including its very high price. They did things like going to MilSpec circuitrt components that could handle higher temps and vibration, tighter tolerances and higher wattages on resistors and a thicker circuit board, all of which was plug and play compatible with the existing GM platform and mounting location. Around the same time Stanadyne came out with their "improved" PMD - the gray beast that uses a slight modification of the plug, necessitating either modifying the harness plug slightly or buying an adapter from - Stanadyne/GM.
The other benefit is that the Flight Services design is not only better than the OEM design, but cheaper in price than either of the Stanadyne designs, too. It wasn't very long after that that FS licensed the product and aftermarkets like Dorman was selling the FS unit under their label name.
Yeah, go ahead and tear one apart for curiosity's sake to see what makes one tick, but don't think you can build the better mousetrap and retire to Disneyworld for the rest of your life, that ain't gonna happen with this engine platform.
