I found an interesting article at The weird, worrisome mystery behind America's plague of purple streetlights that we have seen here

"...It has to mean something. Because if it doesn't, that's even scarier. Streetlights and street lighting are a city's deep infrastructure. If they can break in such a weird and unexpected way, so can everything else."

That is telling everything...

Otherwise as usual: Cheepeese c$$p takes over because bean counters just go for the lowest price tag, not matter what...

That's what I thought, Medved

I love that bit, ‘5G radiation or government surveillance’ 

Don’t even get me started on the tin-foil hat nonsense, the amount of BS is pretty unbelievable… like they have lasers on them now.
 
I wonder which LED manufacturer will earn the guinness world record of most recalled fixtures: Leotek, Acuity, or Cooper? I’ll be quiet.

Don’t even get me started on the tin-foil hat nonsense"

Why? To me this (along the "Flat Earth" and similar) is a great source of amusement and laughter... :-D

Well, they wanted the things; none of this would have happened if people had stuck with proper traditional light sources

Yeah some of the "theories" floating around on the internet regarding these are pretty wild.

The problem seems to be limited to AEL/Acuity luminaires from what I've noticed. My area started using LED in late 2009/2010 and I've yet to see a LED luminaire fail in such manner. Granted AEL/Acuity LED luminaires are rare in my city.

I think the issue lies with whoever supplied the adhesives for the phosphors to Cree. I've seen Autobahns, those CC center span lights, Cree shoeboxes, Lithonia wallpacks, household Cree bulbs, and many more shift to the purple color. They haven't changed the designs during this, but due to supply issues at the time, I'd assume they had to go elsewhere for the material and got faulty stuff. No specific lighting brand is to blame.

No Cree LED lantern here in Israel, has been turned purple.

So you have seen every installation of lighting in the whole country.? 

I think the issue lies with whoever supplied the adhesives for the phosphors to Cree.

There is no dedicated adhesive. The coating itself is supposed to directly adhere on.

The delamination happen, when the surfaces or tge blob mix are contaminated, but mainly when the LED is operated at too high temperature, exposure to humidity (when it is off) worsen it further.
And tge thermal management is the matter of "fixture" design and assembly quality, so "falls on the head" of its manufacturer.

The problem was, that time the LEDs alone were still quite expensive, but there was very strong push to make them "affordable". That led to very high loading of the LEDs in lanterns of a given power, saving on the amount of LEDs needed. It was time when many modern thermal management techniques (tailored for LEDs) were just introduced, but industry had not much experiencd with them yet, leading to frequent design errors and manufacturing quality problems (partly because those designs were not robust enough yet), where tge LEDs got overheated and getting green as the result.

Plus that was the time, when the LEDs alone become covered not by a hard plastic, but by a soft silicone based material. And that meant any pressure onto the LED "lens" directly caused stresses on the bottom side of the material, causing from bond wire breaks (the datasheet did mention this), to crack forming in the interface between the substrate/die and the "lens blob". So when the components were not handled with enough care (that includes setting up tge assembly pick&place machine), those initial crack did form, trap the humidity and with following high heat cycles propagate.

I suppose I should clarify, when I say adhesive I am referring to whatever material there is to help the phosphor stick to the diodes, even if its part of the blob/pad mix itself. I've noticed the phosphors don't just fall off, they crumble apart exposing the emitter at varying degrees, so yes I'd assume not stuck on with a separate bonding agent, and indeed part of the mix as you say.

 It also seems to be affected by heat more then anything. Lights that day-burn or are in shaded areas where a photocell would run them for longer would shift quicker and in a much more aggressive fashion than ones that run for less hours.

The design of the diode arrays and heat sinking however had not been changed at all (at least on the Autobahns) between when the lights had started doing it and before (and on many different designs for the other manufacturers), which is why I think it was a manufacturer defect with the diodes themselves rather than something caused by a design change or overheating.

Also, Leotek still uses the soft rubbery material on the diodes on their lights, however there is a protective cover over them and they do very well even  with rough handling. I haven't had to replace any due to failed diodes out of tens of thousands on the system I work on. It's nice to see that things are advancing with time.

Yeah the LED's turning purple situation is actually starting to get rather crazy. It's so bad, while I was on YouTube, viewing a non lighting related
video of someone out on the road during the night, in his video, all of the LED's were purple/slight blue. This was somewhere in Florida, not
sure the exact location. Said person stopped the vlog to ask his viewers about why the lights were purple, etc.
This purple phenomenon has actually made it to some parts of my area as well, mainly certain empty parking lot areas, some highway LED fixtures...
I was actually considering creating a thread about this, but of course, it has already been done.