12V-2X6 "H++" Standard Touted to Safely Deliver 675 W

[T0@st]

Online hardware communities continue to discuss the 12VHPWR connection standard's troubled existence, while revised technology gets worked on—quietly—in the background. PCI-SIG's 12V-2x6 connector was first revealed last summer, signalling an alternative power delivery method for high-wattage graphics cards. Past TPU reports show that the 12V-2x6 16-pin design has already popped up on select NVIDIA Founders Edition cards, GeForce RTX 40 SUPER custom graphics card designs, and various new generation power supplies. Earlier today, Алексей (AKA wxnod) took to social media and posted an image of the freshly deshrouded "H++" 12V-2x6 (total design limit: 675 W) socket, as well as a shot of the familiar "H+" 12VHPWR (max. 600 W).

This fifth generation socket design largely rolled out with Team Green's GeForce RTX-40 SUPER card series, although wxnod notes that exceptions do exist: "Some AIC GeForce RTX 4070 SUPER, 4070 Ti SUPER and 4080 SUPER cards are still using the H+12VHPWR interface." The H++ identified 12V-2x6 design's power limit peaks at 675 W—a technical breakdown from last July revealed that 75 W comes from the expansion slot, while the big 600 W portion flows through the 16-pin connector. As mentioned before, 12V-2x6 debuted on a few Non-SUPER cards back in 2023, but last month's SUPER series product launch marked a more comprehensive rollout. AMD has indicated that it is considering an adoption of Gen 5 H++ in the future, but we have not heard much on that subject since last August. A new generation 16-pin PCIe 6.0 power connector design was linked to the upcoming NVIDIA RTX 50-series of "Blackwell" GPUs, but Hardware Busters has refuted rumors generated by Moore's Law is Dead. Team Green is expected to remain faithful to "H++" 12V-2x6 with the launch of next generation graphics cards.



View at TechPowerUp Main Site | Source

 

[Shihab]

Another increase?
At this rate, in a couple of years, Americans will have to run their computers off their ovens' sockets. </s>

Seriously, I think it's about time the industry goes for another efficiency-first phase. A few years of Kepler-esque products would do everyone well.

 

[N/A]

The genie is out the bottle. We want more unlimited powaa please. Under volt is still athing and free for all, you can run at 50% power limit and still enjoy 85% of the performance, so. One has the pins just elongated for foolproof that's what the + stands for. And this is based on MicroFit+ enhanced that raises from 10.5 amps to 12.5 and min 16awg that is probably the reason.

 

[FoulOnWhite]

funny how everyone whines about CPU power use, but it's fine to buy a 4090

 

[Onasi]

I am sure someone will still manage to burn their card to ashes with the new connector and the drama will continue.
625 watts on the GPU though… That’s almost comical. What are the coolers for something like that would be like? 5 slots with 3 fans and another on the side, like some Inno3D designs? A Fury-esque AIO?

 

[N/A]

Anyway how can you trust a source that doesn't bother to resize images for side by side up to scale. Lllook how much bigger it is, no it's the same. it mandates the use of 16 AWG cables. What else is different. Any serious PSU maker already does.

 

[mechtech]

675W/12v = 56.25A / 6 circuits = 9.375A per circuit

so ya 16 awg wire

 

[Vayra86]

funny how everyone whines about CPU power use, but it's fine to buy a 4090

Well at least with a 4090 that power pays off

 

[close]

The genie is out the bottle. We want more unlimited powaa please. Under volt is still athing and free for all, you can run at 50% power limit and still enjoy 85% of the performance, so. One has the pins just elongated for foolproof that's what the + stands for. And this is based on MicroFit+ enhanced that raises from 10.5 amps to 12.5 and min 16awg that is probably the reason.

Right. But you're paying through your nose to get a part that's overclocked to the limit of the transistors just so you can then undervolt it for 85% of the performance? The last 15% of performance usually account for 30-50% of the price. On top of paying for the consumed power. As long as people are willing to dish out for home radiators that also render frames, we'll get home radiators.

 

[asdkj1740]

thats bs.
max power consumption comes from two sources, the pcie connector/16pin AND peg connector.
12v2x6(h++) means 16pin=600w + peg=75w only, not new 16pin=675w.
stop copying news plus misunderstanding the source.

 

[LabRat 891]

675W/12v = 56.25A / 6 circuits = 9.375A per circuit

so ya 16 awg wire

So, it's still +12V An Hero, got it.


Unless they've standardized on silver-cadmium new-geometry pins, I'm not seeing how this can safely carry those kinds of currents.
Perhaps, the entire industry has forgotten the safety margins in EPS and PCI-e power standards?

I am sure someone will still manage to burn their card to ashes with the new connector and the drama will continue.
625 watts on the GPU though… That’s almost comical. What are the coolers for something like that would be like? 5 slots with 3 fans and another on the side, like some Inno3D designs? A Fury-esque AIO?

To give some perspective. 625W is roughly half the max-rated output of an HP Common Slot redundant PSU module.

That PSU is about the size of a ' 2"x4" ' stud, and the fingers for GND and +12V are double-sided.
Additionally, the 'finger' style acts as a heat-sinking plane for any variations in contact-to-contact resistance.

My point:
At no time, at no place, for no application,
has such a glaringly lacking 'safety margin' been acceptable. Basic observed physics/thermodynamics + observations of past high-current +12V interfaces, shows there's something fundamentally wrong about +12V2x6/HPWR

 

[HisDivineOrder]

This feels like the kind of thing Nvidia should have worked on way before the current release rather than keep releasing beta cords (of all things) on repeat.

Or they could just not put that much power through a gaming card.

 

[maxfly]

5090 BZZZTT ti Super White Lightning

Coming Q4 24!
Buy two and save, save, save!

12VHPWR ver.10 cable included free with each card.

 

[Why_Me]

This is great news.

 

[FoulOnWhite]

Well at least with a 4090 that power pays off

Unless your card melts itself, not such a pay off then eh

 

[user556]

They're all staying real quiet on the cause of the melted plugs.

 

[Vayra86]

So, it's still +12V An Hero, got it.

View attachment 335424
Unless they've standardized on silver-cadmium new-geometry pins, I'm not seeing how this can safely carry those kinds of currents.
Perhaps, the entire industry has forgotten the safety margins in EPS and PCI-e power standards?


To give some perspective. 625W is roughly half the max-rated output of an HP Common Slot redundant PSU module.
View attachment 335427
That PSU is about the size of a ' 2"x4" ' stud, and the fingers for GND and +12V are double-sided.
Additionally, the 'finger' style acts as a heat-sinking plane for any variations in contact-to-contact resistance.

My point:
At no time, at no place, for no application,
has such a glaringly lacking 'safety margin' been acceptable. Basic observed physics/thermodynamics + observations of past high-current +12V interfaces, shows there's something fundamentally wrong about +12V2x6/HPWR

Perhaps they're trying to realize a new definition of 'you'll own nothing and be happy'? Just burn that shit down

 

[FoulOnWhite]

Just don't buy a card with one of these connectors. If they sell none they will have to sort it out, but i guess it's too late now, too many already out there. Looks like this type of connector is here to stay.

 

[trsttte]

Just don't buy a card with one of these connectors. If they sell none they will have to sort it out, but i guess it's too late now, too many already out there. Looks like this type of connector is here to stay.

Not necessarily, AMD is having none of it and probably so is Intel. Contrary to nvidia, they allow board partners some design freedom and given the option they won't have any of this shit either (the older connector is cheaper, easy choice really). Server side never gave a fuck - they've often used AT connectors instead of PCIe spec anyway - and I don't see them starting now. Only ingredient missing is really PSU manufacturers refusing to continue to go along with this stupidy - kind of surprised they ever went along with it really.

At some point all of these factors will put pressure on pcie-sig and nvidia to fix - or rather un-screw - the situation.

 

[close]

Not necessarily, AMD is having none of it and probably so is Intel. Contrary to nvidia, they allow board partners some design freedom and given the option they won't have any of this shit either (the older connector is cheaper, easy choice really). Server side never gave a fuck - they've often used AT connectors instead of PCIe spec anyway - and I don't see them starting now. Only ingredient missing is really PSU manufacturers refusing to continue to go along with this stupidy - kind of surprised they ever went along with it really.

At some point all of these factors will put pressure on pcie-sig and nvidia to fix - or rather un-screw - the situation.

People who dish out for a 600W GPU are likely to go for high-end (expensive) PSUs so I can't imagine PSU manufacturers will want to turn their backs to their halo-product buyers. Not because those bring the most profit (they don't) but they carry the brand name forward.