It’s on Aliexpress. The very CREE that I used in comparison was the 90+ CRI 5000k version bought from HotRed Store. The very same store also sells 6500k and 3000k variants. You get to choose voltage, too.
uhh, nevermind, the 90CRI version only available there in 5000k.
You might try your luck here, but it’s 1000pcs minimum order.
Update:
The manufacturer says the CRI is indeed 95, and they won’t make the label for 95 CRI.
They have yet to confirm if they can produce the A4 Bin. So, don’t hold your breath for that.
Nevermind, they said they can do A4 or even C2, if there are enough people interested in.
Why do you call it a bigger version of XHP35 HI? Why not just say it’s like a XHP-70 HI? Anyway, I’m eager to see what the “collimated” beam from these look like compared to the XHP-70 when you finish the Sofirn C8 builds. Will you also de-dome the XHP-70 to compare with both LEDs being domeless? 
This sounds exciting. If everything test out well, perhaps we can get MTN or Kaidomain interested to buy in the quantities they need and we can just buy from them.
The reason I was reluctant to call that a XHP70 HI is that the given spec wasn’t as efficient as the XHP70.2. By estimate it should be about 20lm/w behind at the same CRI.
By the way, just want to let you know that I also plan to later buy these in bulk to make part of the living, and also gain funding to hunt for another product.
No problem with the reduced efficiency. We expect that with all high CRI variants. If you check CREE specs, you can often find that their high CRI versions of emitters are at least one or two brightness bins lower than their low CRI version of the same emitter. In other words, the high CRI version just isn’t available in the highest brightness bins. Same with low CCT (warm) versions. They can’t make them as bright as higher CCT (colder) versions, so the highest brightness bins just aren’t available in the low CCT versions of their emitters.
I see I underestimated the lm/w lost going from 90 to 95 CRI, so let’s see the calculation:
Based on GeTian’s previous product I sampled:
Going from 70 to 90 CRI - you lose 20lm/w
Going from 90 to 95 CRI - you lose another 20lm/w
And based on CREE’s datasheet:
Going from 70 to 90 CRI will cost you 21-25lm/w
Assuming that you lose the same 20lm/w for CREE’s to go from 90 to 95 CRI:
You would look at 70-90 lm/w for the theoretical 95+ CRI version of CREE’s XHP70.2 (or 2200-2900 lm at maximum rated power)
What would be good drivers/flashlights for these LEDs? I only have 3v drivers/flashlights atm so don’t know much about others.
Looking forward to the tests!
I would be very interested in C4 but have a feeling A4 will be more popular.
The readily available driver are the ones that used to drive the XHP35’s, but it’s only driving the LED at 1A so you would be seriously underdrive these LED (though not necessarily a bad thing).
On the other hand, these are 40w LEDs so running these on full tilt in flashlights with poor thermal dissipation could very well be a recipe for disaster.
Shouldn’t you wait for Djozz test results before posting the groupbuy? Most people are not going to buy without seeing what they are getting.
I should be able to at least do a proper beamshot and side-by-side comparison against CREE’s on a flashlight this week, provided I won’t destroy the driver.
I could always let Djozz chime in later. Djozz’s result should turn up next two weeks.
Also, the manufacturer seems quite confident about CRI values of this thing. They did said something along the line of “see it for yourself”.
I have no doubt it will reach 95CRI however I’m more interested in the tint at different currents and output at different currents.
I guess this week’s beam shot should serve as some kind of appetizer on the tint and output on one specific current. 
I have both of them in a C8s host today.
I got beamshots too. Unfortunately, I only have one high drain 18650, and my setup trips the overcurrent protection of other batteries. I can only give you the individual beamshot for the moment.
All the shots are taken at 5500k WB. I took this with a smartphone camera which annoyingly messes with the exposure setting.
XHP70’s Beam shot
FC40’s Beam shot — Again, This looks much less green than XHP70’s 90CRI, but a bit cooler.
Additionally, the H2-C I ordered are not the same. One of them operates in 1-mode, the other operates in 5-mode. Annoying, but it’s only the best shot I have at this. I probably don’t want to go through the trouble of sanding down the driver and coat it with nail polish again.
Anyway, let’s light up an entire room.
The tint looks great compared with the greenish XHP70.2. Let’s see how output and efficiency test out.
I’ll wait for the test result but sound promising…see the beamshot…I don’t know if this is “going to be next new”…wow, 
I did some very crude brightness measurement by shining the hotspot on a luxmeter at 1ft distance.
The measurement reads ~120k-140k lux on the highest mode (with the H2-C)
That equals to 11k-13kcd or 210-220m of throw distance in a C8s body
I’m now searching for the proper 20mm driver for these, but it’s really hard to find.
11k-13kcd in a C8 head seems very low.
That could be two things: