Does not look fake to me. Beides, many companies plan to make those including VS/Panasonic
Magnetic ballasts can be made up to any desired efficiency - just use better materials for the coil (thicker wire) and core (better metal, more plates and each one thinner etc). Want 99% ? No problem - mage good big core and wind it with 10AWG. Only limitd by cost (size and weight i dont count since there is plenty of space in any fixture, so its not issue)
The "Energy efficiency label" is not supposed to compare the efficiency (as per engineering/technical/scientific definition), but take into account other features potentially affecting the overall energy usage.
With fluorescent ballasts it is the high frequency operation (based on the definition, the standard count 15% lamp efficacy improvement for the HF drive compare to the mains frequency) or the capability to reduce the lamp power to actual need, so it is supposed (according to the standard operation) does not have to be at full power at all times, so save the power there.
(And e.g. with cloth tumble dryers the difference between "C" and "B" is the "B" requiring a single-push-button mode for low load, even when the same effect you may get by proper setting of the simpler control of the "C" dryers - it is not single push button, so it can not be "B", adding the ice making capability into freezer mean it belong to different appliance category, so although the ice maker is unusable in the central Europe and further to nord and only consume extra power, such appliance could usually shift one degree to better grade)
Well, this make the "energy efficiency" label system a true practically unusable mess, good enough to only fool potential customers...
With fluorescent ballasts A1 is less efficient than A2 and A2 is actually less efficient than B1 (speaking of energy efficiency in converting mains power to the power for the lamp, not about system efficacy in converting mains power into useful light).
The "B2" stand for mains frequency ballast of standard efficiency (~20% of losses)
The "B1" stand for mains frequency ballast of high efficiency (<10% losses). When you don't mind the heavy and expensive control gear, there are the most reliable solution: It is still robust coil and core, while the low losses and large size ensure very cold operation.
The "A2" stand for a HF (so electronic) ballast without dimming, losses up to ~15% range (I don't remember exactly; they are about 10% on most ballasts, better for the simple instant start ones, worse for the fully protected programmed start models, the high PF ones have higher losses than normal PF ones, but with the large installation the extra losses in the wiring system could be actually higher than losses of the PFC). Because of the high frequency drive the lamps are expected to be ~10..15% more efficient than on magnetic ballasts, what allow energy saving even (with higher losses) than "B".
But the "extra losses" could be sometimes so high, than the system actually become on par with more efficient (<5% losses) models of the "B1". Well, in real life with many tubes (mainly the thin, high current density ones) the high frequency does not improve the lamp efficacy as much, so the lower B1 losses could actually mean ~10% better system efficacy than A2...
The "A1" require a dimming feature capable to lower the power usage when only partial output is required. The extra net energy saving against A2 is expected from the use of this feature, even when the incorporation of dimming feature mean extra losses in the system (so lower efficiency compare to A2). Again in real life are not as many places, where the dimming feature could really be efficiently used to save power (against e.g. switching half of lamps OFF in an installations with two lamps per fixture or when no light reduction is possible at all), so in most cases the extra losses of the dimming circuitry in A1 mean actually higher energy usage compare to A2 (and sometimes even B1)
So if the cited ballasts are really magnetic ballasts, they are B1, because with the hypothetical reference lamp it won't utilize the hypothetical 15% gain in efficacy.
And the high voltage nature of the T5HE mean, than regular glowbottles won't work. The ballast may stabilize the arc by using the serial capacitor, so the arc voltage could be as high as 80% of the mains (the resonance effect between the L and C make sure, the tube have up to 2x the mains voltage for reignition after zero cross), but when there is not yet any current, the OCV is still equal to the mains. So the voltage across the tube is not usable anymore to determine if the lamp had started (so the starter have to "step away") or not yet (so the starter have to kick in). What remain is the time (where the starter change actions purely based on the time since power ON, without sensing what the lamp actually does; 80W T5HO worked in this setup well with the S10E electronic starter with "F36T8" ballast and 3uF/500VAC in series, the starter changed modes "Preheat" for 2.5seconds->"Ignition attempt" for 1 second->"Fail Shut Down", where the lamp in fact remained lit after that). Other option is to use the circuit current as a signature of running/not started lamp (e.g. like Perfect Start thermal starter system).