I'd not say they really have, they run single train pairs on most of their routes, and will have an impact on DB no greater - probably lesser - than the effect something like Lumo or Hull has on LNER.
The biggest impact they've had is locally, because currently, track access in Germany is still based on dreams of a magic free-for-all-competition wonderland, resulting in an infrastructure operator being required by law to give priority to those train paths that net the highest income. This means that in the case of conflicting train paths, regional trains usually get the short straw and have to get out the way, thereby breaking clockface timetables and scheduled connections. (While there is a provision in the law that "Takt"-train paths get priority, so-called "network integrated" trains are treated in equal terms, and the criteria for that latter category aren't too difficult to fulfil even for open access operators running only one or two trains a day on a route.)
So basically everywhere Flixtrain decides to run, it leaves behind a small bloodbath of broken clockface timetables and connections, or some trains (or at least some stops) cancelled completely, with passengers on those trains having to travel one train earlier or later, whenever that may be. To be fair, even DB itself isn't perfectly innocent with scheduling their long distance trains in that regard, but basically every additional open access operator exacerbates these kinds of problems, so from that point of view we can be thankful that competition among long distance trains hasn't really taken off so far.
Some years ago the law was modified to allow an infrastructure operator to unconditionally give priority to clockface regional services, but it was an opt-in provision and DB Netz never decided to take it up, so it never made any difference in practice. The rules for dealing with sections of the network that are formally at capacity (i.e. where not all requests for train paths could be met) have also been changed in a similar vein, but likewise as far as I know those new provisions have never actually been applied in practice so far.
Finally, quite recently DB Netz itself proposed a concept ("mKoK") whereby a globally, network-wide optimised timetable is being developed based on known flows of traffic (with operators + the state authorities responsible for regional trains being able to give their input in that regard) and then subsequently train paths that are bid according to that sample timetable will get priority over train paths that deviate from that timetable. This would at least solve the problem of uncontrolled open access competition being currently able to randomly wreck clockface timetables and connections (though it still leaves the problem of through-ticketing – by law railways are actually required to work together in creating through fares, but in practice as far as long distance trains are concerned that section of the law is basically being ignored), but unfortunately shortly after the initial announcement DB Netz had to back-pedal and announce that that new concept would have to be delayed by a few years because it turned out that introducing such a thing from scratch was more complex than expected.
Interesting view on the far older and less efficient electrification system (15KV 16.66hz OLE)
Before the advent of three-phase drives + AC/AC converters and the resultant ability to keep the power factor near 1.0, the higher voltage (25 vs 15 kV) only mostly compensated for the higher inductive impedance caused by the higher frequency (16.7 vs. 50 Hz), so all in all it wasn't really more efficient. Plus the requirement to feed adjacent sections from differing phases in order not to overload the national grid with unbalanced loads results in frequent neutral sections and each section being fed only from a single point (meaning higher voltage drops when moving away from the feed point), thereby again driving down section length and negating the advantages of the higher voltage.
So if you're starting from scratch, 25 kV 50 Hz decentrally fed from individual substations is a bit easier and cheaper (and nowadays substations based on solid-state inverters instead of simple transformers would also allow both avoiding the problem of unbalanced loads on the national grid as well as getting rid of the requirement for neutral sections, though they're still a bit more expensive than transformers, so you have to weigh those advantages against those additional costs), but I came across some study on that subject which concluded that on the other hand in those 15 kV 16.7 Hz countries with an already existing separate railway power grid, keeping the existing railway power grid is more advantageous as compared to attempting to slowly getting rid of it.
and much more complicated layouts (with a fondness for double slips).
Dunno, looking at a number of major UK stations I wouldn't call those layouts non-complex, either. I also don't think Germany ever really had that diamond crossing madness which e.g. used to exist at York or Newcastle and still survives in a smaller form at e.g. Shrewsbury. And in Germany the smallest set of points still allows 25 mph/40 kph (Cologne with its 30 kph speed restriction on the eastern station throat is I think more or less the only exception), while in the UK points only allowing 15 mph or so aren't too uncommon.
After RENFE lost it's decade long monopoly on operating the Madrid - Barcelona line, prices dropped by 50% and ridership doubled in just a few months.
Open Access operators like FlixTrain and Nightjet serve viable markets the incumbent operators just don't care about, yet Trenitalia, DB and most egregiously SNCF have all been caught out playing dirty tricks to try and make life difficult for 3rd party operators.
That's all very nice when you a) have large point-to-point flows and can ignore connecting trains (in France of course all flows are dominated by Paris, in Spain likewise the population is somewhat more concentrated in the bigger cities and less spread out, plus RENFE already mostly doesn't seriously care about connecting trains what with its default minimum connecting time of one hour (!) and so open access competitors not offering connections don't really make things worse) and b) your network has plenty of spare capacity and ideally those open access operators can keep themselves to those shiny new high speed lines (the Spanish HGV network isn't exactly known for being overloaded for example).
On a more heavily used mixed traffic railway network the idea of free competition however rapidly breaks down. First of all this means that instead of directly competing for passengers, operators start competing for train paths instead, which means that incentives between the railways and passengers are already starting to get dis-aligned, and secondly its e.g. incompatible with having regular interval timetables and scheduled connections. When a regional service e.g. only runs hourly (or even if it runs half-hourly), it can basically only have good connections (not too short, but definitively not too long, either) with only one or at most two sets of long distance trains per direction, so who gets to decide which company gets to have that good connection, and which one doesn't? Ticketing incompatibilities then are the cherry on top of all that.
The TGV Duplex may get more, but it has a signficantly tighter seating arrangement than an ICE, and you could do that on the single-deck train.
… and it has noticeably worse dwell times – not so much of a problem with the French model of dominantly running point-to-point services to Paris and back again, but less tenable with the kinds services typical of Germany which often have major intermediate stops with a large fraction of passenger getting on and off.
