You need three things to have a fire - a fuel source, an oxidizer and an ignition source. O2 is an obvious oxidizer and most materials will burn in a pure O2 environment given a hgh enough temperature (ignition source), so it is all pretty much a matter of degree.
The approach to a "safe" nitrox reg is done on three fronts:
1. Limiting fuel by using "O2 compatible" materials such as viton and EPDM o-rings. Both have higher ignition temps thab nitril rubber o-rings. Viton has a ligher ignition temp than EPDM, but does not wear as well in service, particularly in dynamic applications. O2 cleaning to remove contaminants is alos intended to remove potential fuel inside the regulator. Titanium is problematic in this regard as the oxide that forms on it is a potential fuel, so they are not well suited to use with nitrox over 40%.
2. Limiting 02 percentages to limit the oxidizer present. This is where the issue gets controversial as people have argued for well over a decade on how much is too much. The scientific theory involved in the argument is complicated by economics. Nitrox was adopted by the mainstream dive industry after years of condemning it when the industry finally realized how much profit there was in it for training agencies as well as manufacturers. In that regard the moral high ground some still try to stand on with a 23.5% standard is a bit tainted given that O2 cleaning is a revenue source. In any event, theory is what it is but from a practical standpoint, the 40% standard seems to work fine.
3. Limiting ignition sources. This one is more difficult in that valves and regulators for scuba are not well designed for O2 service. In general, scuba valves open quickly with rapid rises in pressure. That combined with dead spots in the flow path, or sudden stops at the end of the path (such as a regulator with closed second stage valves) can create rapid compression and that generates heat. Rough spots or sharp corners in the flow path can also create hot spots. In contrast Medical O2 and welding regulators and valves are designed to pressurize comparatively slowly and have flow paths designed for O2 use. To reduce the temperatures and the potential for ignition, nitrox tanks and in particular tanks exposed to O2 over 40% are filled very slowly to limit heating during compression and reduce velocities in the flow paths to reduce heating as well as to reduce the potential for a spark from an entrained metalic particle.
A few loose ends:
O2 does not "explode", it just burns fast and cause a rapid rise in pressure - not an explosion.
Too many nitrox divers seem ignorant of proper technique. To reduce compressive heating in the reg, it is best to pressurize it slowly while having the purge button slightly depressed - not the far more common spinning the valve open with the reg crreating a closed circuit.
Tanks are not the major problem in terms of O2 clean as it's hard to introduce a high ignition temp into one. People get all upset over silicone lube on the valve threads but it would be difficult to get that to reach it's flash point anyway. A clean valve is far more critical than a clean tank as there is greater potential for hot spots and ignition inside the valve. It is also exposed to the O2 for long periods of time making materials that do not degrade in the presence of high O2 percentages more important.