Many, I think, mostly maintenance related. Frequent causes involve old gaskets or cracked fuel lines dripping onto the exhaust manifold, right on the side of the engine.
You're not wrong that fuel does sometimes catch fire but it is very rare for a fuel leak to cause a fire on its own. It usually happens after a fire starts elsewhere.
Fuel system integrity in crashes has been scored in crash tests for a very, very long time (I recall reading about it when I was reviewing some microfiche with crash tests results of 70s pickups) so OEMs have been trying to build secure fuel systems for a long time and have gotten pretty good. With cross flow cylinder head designs (which every engine today has) there is no reason (it would be a waste of material to plumb it) to have fuel on the same side of the engine as the exhaust. The routing between the engine and the vehicle body basically always (i.e. I know of 0 exceptions but can't prove one doesn't exist) go out of their way not to run over exhaust. Fuel systems rarely have gasketed connections, almost everything is sealed with O-rings.
Mature EV designs probably catch on fire less than ICE designs but not really for any of the reasons you specified. Fuel fires mostly come from crash damage to fuel tanks where the crash also leaves the vehicle overturned. Most fires come from electrical systems not related to the drive-train. Even engine fires are rarely fuel, animal nesting debris is a far more common cause.
Ford recalled 8.7 million cars for spontaneous fire risk.[1] And that's just one case.
It happens more often than you'd think, and coverage of it in media tends to be subdued. Automotive is the largest source of advertising dollars for many media markets.
[1] https://www.latimes.com/archives/la-xpm-1996-04-26-mn-63066-...
2,000 cars in the U.S. and Canada were believed to have had fires related to the ignition switches, which were installed in 23 million Ford vehicles. This is a failure rate of about 0.0087% overall, although Ford ultimately argued that it would only need to recall a subset of the vehicles with the ignition switch: 8.7M vehicles.
There are at least a dozen [1] Tesla fire incidents, and 720,000 Tesla vehicles sold through 2019 Q2 [2], which is a failure rate of about 0.0014%. So Tesla has a lower rate, but not by an order of magnitude, because Tesla has sold far fewer vehicles than Ford did over roughly the same number of years. So it seems justified for NHTSA to take a serious look into the issue.
[0] https://www.washingtonpost.com/archive/politics/1996/04/26/f...
[1] https://en.wikipedia.org/wiki/Plug-in_electric_vehicle_fire_...
How many modern vehicles spontaneously burst into flame? I could easily believe it's orders of magnitude larger than for Teslas, but when you make comparisons like you did it just discredits your entire argument.
Oh, and the majority of present-day ignitions in automobiles are still the exact same mechanical interlock switching types used since the 60's. Higher-end vehicles have solid state switching, but it's not the majority of vehicles.
Also:
https://abcnews.go.com/US/bmw-recalls-million-vehicles-fire-...
The burden of proof is on the person making the claim.
https://www.logicallyfallacious.com/tools/lp/Bo/LogicalFalla...
Anyway, agree to disagree I guess.
Of course it's still rare, and I'll also mention that the Model 3 has a newer battery pack design and is not being investigated.