• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

Is your 5000 series nVidia card efficient? Idle power draw comparison (5070/5070Ti/5080/5090 & others)

Joined
Apr 26, 2025
Messages
63 (0.85/day)
So, for some time I've been trying to collect data on the practical idle power draws of the Blackwell-series cards. Why do I care? Because that's power consumption that's being used any time the computer is on, even if there is no load and the user is AFK. This warms up your case and can cause higher fan noise not just from your GPU, but also case and even CPU fans. It also warms up your room and (slightly) racks up your electricity bill.

TPU has good data for a lot of cards, but I'm noticing that most of the cards reviewed are the higher-end models. I'm presuming most AIBs send in their best, since they want to put their best foot forward when launching a new product.

However, GPUs are already too expensive and the only way to fight back it to look at models that are being sold closer to MSRP.

For this reason, I spent a few days asking people that I know have 5000 series nVidia cards about their idle power. This is what I've found. Each datapoint is a single GPU of that model. If you see multiple - those are different cards. Unless stated otherwise, the OS is windows. I've noted the type and number of monitors for those who shared. If it isn't listed - don't automatically assume single-monitor, I just don't know.

  • 5090
    • PNY - 50W (multi-monitor; Linux)
  • 5080
    • PNY - 39W (multi-monitor)
    • PNY - 35W (multi-monitor)
    • MSI Ventus 3X OC Plus - 30W
    • MSI Shadow 3X - 31W (multi-monitor; first monitor 4k 120hz, second monitor 1440p 165hz)
  • 5070Ti
    • MSI Shadow 3X - 15W (single-monitor); 35W (multi-monitor; 3x 4k monitors)
    • MSI Gaming Trio - 45W
    • MSI Ventus 5070Ti - 18/19W

I've also done some ad hoc aggregation of TPU data and reviews from other YouTubers, including some foreign-language reviews. What I've found is that Palit seems to have a relatively high idle power draw, followed by Gigabyte. MSI at times appears to have good power draw, but some posts on reddit state they have terrible power draw (40+ watts on a 5070Ti). Palit, Gigabyte and MSI are at the moment releasing some of the most affordable GPUs, so it would be great to get more data on them. Palit has consistently been pushing for MSRP in Europe. Inno3D cards also appear to have a very low number of reviews. I did find one review placing it at just over 20Watts for a 5070 Ti.

How does your card compare? The cause demands that you share your idle power draw. If you do, it is also helpful if you mention how many monitors (even better if you mention what type) you have running. Also, posting your opinion on the cards in general (are they loud? hot? anything that annoys you?) is also very useful.

Share your idle power draw. You know you want to. It is time for this data to reach the world. Think of it as your contribution to fighting climate change. You are a warrior of science, gaming and ecology.
 
Palit 5090 GameRock - 1x 4K/120 G-sync display, only opera browser with this thread open. Windows 10 Pro 22H2 576.15 driver

32 W

gpuz.gif
 
MSI Suprim SOC.

Board power ~12 to 13 W. 4K 240Hz monitor.

gpu-z 4-27-25.gif
 
Colorful Tomahawk RTX 5060 Ti DUO 16 GB (or known as the Battleax here)
8.9W board power at the lowest, though occasionally spikes to 20W due to things drawing on my desktop. Edge and Discord open, as well as the NVIDIA app performance monitoring overlay... so not the absolute lowest?

Two displays, 2560x1080 at 200Hz, and 1920x1080 at 60Hz
 

Attachments

  • Screenshot 2025-04-27 222144.png
    Screenshot 2025-04-27 222144.png
    160.5 KB · Views: 123
gigabyte windforce 5070 ti (with aorus 400W vbios) idles at 25W with two 4k monitors.
Damn, how's that card holding up with the 400W vbios? How much did you crank the power limit?

Palit 5090 GameRock - 1x 4K/120 G-sync display, only opera browser with this thread open. Windows 10 Pro 22H2 576.15 driver

32 W

View attachment 397072

That actually seems great for a 5090! Beats the PNYs I polled and the TPU models.

MSI Suprim SOC.

Board power ~12 to 13 W. 4K 240Hz monitor.

View attachment 397131

5080 at 12-13W at 4K @240hz is amazing! That's a definite winner card. Gj bro!

Colorful Tomahawk RTX 5060 Ti DUO 16 GB (or known as the Battleax here)
8.9W board power at the lowest, though occasionally spikes to 20W due to things drawing on my desktop. Edge and Discord open, as well as the NVIDIA app performance monitoring overlay... so not the absolute lowest?

Two displays, 2560x1080 at 200Hz, and 1920x1080 at 60Hz

Seems to be a pretty good multi-monitor power draw, especially considering one is a high-refresh rate. Occasional spikes are probably not that big of a deal, it happens sometimes.
 
Only web browser and file manager open. Zotac Solid OC 5070Ti.
 

Attachments

  • gpu-z zotac 5070ti solid oc.gif
    gpu-z zotac 5070ti solid oc.gif
    16.3 KB · Views: 232
  • gpu-z zotac 5070ti solid oc 2.gif
    gpu-z zotac 5070ti solid oc 2.gif
    15.8 KB · Views: 220
I managed to find someone with a 5070 yesterday, so adding this datapoint:

MSI Shadow 5070 2X: 3W at idle when in P8 state and used as a server under Linux. He actually couldn't remember if he had plugged it into a monitor, but on my nVidia GPU p8 shows when plugged in, so I'm assuming he's running it single-monitor.

Only web browser and file manager open. Zotac Solid OC 5070Ti.

Damn. That's pretty high for a 70ti. Have you considered undervolting to see if you can squeeze some more efficiency?
 
I hope you're aware that the GPU's self-reported power consumption numbers are not accurate. As the TPU reviewer explains,
All power consumption numbers reported on this page are "card only" values measured via the PCI Express power connector(s) and PCI Express bus slot. Everything is measured on the DC side, it's not the power consumption of the whole system. We conduct a physical measurement using professional lab equipment, the values are not software sensor readings, which are much less accurate.
 
Single monitor

Windows 11

MSI SUPRIM Liquid 5090

This is me just browsing the net, downloading some drivers, just living my life outside of gaming.

Symbols are Min, Avg, Max

GenericLogViewer_mbmgOi2Y8S.png
 
"This warms up your case and can cause higher fan noise not just from your GPU, but also case and even CPU fans. It also warms up your room and (slightly) racks up your electricity bill."
Then you porbably shouldn't sit in the room with your PC, because I'm pretty sure YOU will have more of an affect on the room temp than an idling gpu.


Right now my GPU is sitting at 28c(fans off) and my chipset is 36c( no fans except chassis), so I think a chipset is a "bigger" problem than an idling GPU.
 
Then you porbably shouldn't sit in the room with your PC, because I'm pretty sure YOU will have more of an affect on the room temp than an idling gpu.
Then you probably also should split in two and send you [the heat generator] to another room while you [the gamer/listener] stay where you are.
 
I hope you're aware that the GPU's self-reported power consumption numbers are not accurate. As the TPU reviewer explains,
Еven so, with a large enough sample size we ought to be able to compare the different GPU's power draw relative to each-other. If we see certain models, or brands, consistently draw more than others, that is a decent indicator.


Single monitor

Windows 11

MSI SUPRIM Liquid 5090

This is me just browsing the net, downloading some drivers, just living my life outside of gaming.

Symbols are Min, Avg, Max

GenericLogViewer_mbmgOi2Y8S.png
~21 is actually great for a 5090! Less than half of some of the competitors.


"This warms up your case and can cause higher fan noise not just from your GPU, but also case and even CPU fans. It also warms up your room and (slightly) racks up your electricity bill."
Then you porbably shouldn't sit in the room with your PC, because I'm pretty sure YOU will have more of an affect on the room temp than an idling gpu.


Right now my GPU is sitting at 28c(fans off) and my chipset is 36c( no fans except chassis), so I think a chipset is a "bigger" problem than an idling GPU.
True, but I can't stop my metabolism (well, I can, but that would come with unfortunate side effects that would prevent me from utilizing my GPU). I can, however, buy efficient equipment to minimize unnecessary heat output.

Do keep in mind that your GPU probably has a far larger cooler than your chipset's heatsink and hence the effect of passive cooling will be larger. For this reason, power draw, as opposed to temperature is imo a better measurement of how much heat a component is pumping into your room, since it takes the efficiency of the cooling solution out of the equation.
 
True, but I can't stop my metabolism (well, I can, but that would come with unfortunate side effects that would prevent me from utilizing my GPU). I can, however, buy efficient equipment to minimize unnecessary heat output.

Do keep in mind that your GPU probably has a far larger cooler than your chipset's heatsink and hence the effect of passive cooling will be larger. For this reason, power draw, as opposed to temperature is imo a better measurement of how much heat a component is pumping into your room, since it takes the efficiency of the cooling solution out of the equation.
You aren't trying to utilize your GPU, so the side effects wouldn't matter.

Yes, which matters when a GPU is running above idle.
My PC has 7 temps showing a difference of about 2 to 4 degrees.
Ranging from a hdd, to the Ram, to the GPU, all at idle.
If I put the fans on the GPU, it is sitting tied with 3 other components at 24c.
 
MSI 5080 Vanguard , single 1440p @ 165Hz ~16W

5080pw.jpg
 
kinda offtopic, but in stalker2 using framegen and 1440p ultra settings / fps limit 158 the 5080 uses around ~160-180W to keep the framerate around 80-158
my old 3080Ti used ~250-300W and ~60-90fps at same level. seems like magic having almost double the fps at half the power.
 
I scoured around the internet and found some foreign language reviews. This has some interesting numbers on a few cards. Unfortunately, I don't have data relating to the methodology, so I don't know how many monitors, or at what resolutions they were running atl.

Inno3D 5070 Twin x2 (usually the cheapest 5070) - 5.8W
ASUS Prime 5070 OC - 7.1W
nVidia 5070 Founders Edition - 8W
Gigabyte 5070 Gaming OC and the MSI 5070 Gaming Trio both come in at 10W

The MSI Ventus 3X 5070 Ti - 9.7W
Inno3D 5070 Ti - 14.1W
Gigabyte 5070 TI Gaming OC - 25.5W
MSI 5070 Ti Gaming Trio - 23.5W

Inno3D 5090 ICHILL Frostbite - 11.7W (outstanding for a 5090!)

And just for comparison a few other cards:

MSI 4070 Ti Super Ventus 3x - 7.4W

AMD RAdeon 7800XT - 14.9W
AMD Radeon 7900XT - 15.2W (Assuming this is the reference/MBA model)
AMD 7900XTX - 18.5
ASUS TUF Radeon RX 9070 XT OC - 15W
ASUS TUF Radeon 9070 OC - 15.7 (weird that this idles above the XT)

kinda offtopic, but in stalker2 using framegen and 1440p ultra settings / fps limit 158 the 5080 uses around ~160-180W to keep the framerate around 80-158
my old 3080Ti used ~250-300W and ~60-90fps at same level. seems like magic having almost double the fps at half the power.
Being able to run games without the GPU sounding like a jet engine and heating my room up like a space heater is one of the things I look forward to the most when buying new GPUs, especially in the summer. Really allows me to focus on the game and not have to get annoyed by heat or noise.
 
My PNY 5070 (VCG507012TFXPB1-O) reports only 5-6w when idle with a single 1920x1080 240Hz monitor. This is while using the pc too. Browsing the web, email, youtube etc (browser hardware accel off).

When idle with the monitor off (not sleep) hwinfo has recorded as low as 0.886w. And as you can see in the screenshot, over the past 45 hours it has averaged 2.7w. The pc was idle for the majority of it, with the monitor off. But also a few hours of light load (WoW).

For reference, the MSI 3060 (VENTUS 3X OC LHR) that I upgraded from idled around 10-11w and the lowest hwinfo ever saw was 6.8w. In the same exact rig, same windows install.

For anyone looking to get one of these cards, check your local walmarts gpu stocks. I live in a very low population state and found this card at my local walmart 2 weeks ago for $549.

5070.png
 
Palit 4090 = 27-29 watts on a 4k/240 display.

Nvm just checked again with browser and steam off, 21-22 watts. Dropping to 4k/60fps goes down to 15-16 watts
 
Last edited:
GPU:- Palit 5070Ti Gamerock (Non-OC Model)
Monitors:- Main - 34" MSI Mag341CQ (UWQHD) | Second - 24" Asus VX24A
Board Power Draw:- ~30W
1748611919821.png
 
I got my hands on an Asus Prime 5070Ti. Unfortunately, it seems to have pretty poor idle power consumption at 35-40W at idle. Sometimes it goes even higher. This appears to contradict some of the other numbers I've seen for Asus cards, as well as TPU's TUF review, so I'm wondering if anyone has any ideas what could be causing it to idle so high?



--------------------------------------------------------
I've found also that Gigabyte's Eagle models generally have about the same consumption as the Windforce.

------------------------------------------------------------
Seperately, another person I polled told me this:

MSI RTX 5090 Suprim - triple monitor setup, 1 of which being a TV. 68w with all on, 58w with TV off.




My PNY 5070 (VCG507012TFXPB1-O) reports only 5-6w when idle with a single 1920x1080 240Hz monitor. This is while using the pc too. Browsing the web, email, youtube etc (browser hardware accel off).

When idle with the monitor off (not sleep) hwinfo has recorded as low as 0.886w. And as you can see in the screenshot, over the past 45 hours it has averaged 2.7w. The pc was idle for the majority of it, with the monitor off. But also a few hours of light load (WoW).

For reference, the MSI 3060 (VENTUS 3X OC LHR) that I upgraded from idled around 10-11w and the lowest hwinfo ever saw was 6.8w. In the same exact rig, same windows install.

For anyone looking to get one of these cards, check your local walmarts gpu stocks. I live in a very low population state and found this card at my local walmart 2 weeks ago for $549.

View attachment 401632

This is amazing idle power! This is what I want to see. Props to PNY, that appears spectacular. Does it go up much during media consumption (eg. YouTube videos)?


Palit 4090 = 27-29 watts on a 4k/240 display.

Nvm just checked again with browser and steam off, 21-22 watts. Dropping to 4k/60fps goes down to 15-16 watts
That's pretty damn good for a 4090. Many 5070 Tis have higher idle than that.


GPU:- Palit 5070Ti Gamerock (Non-OC Model)
Monitors:- Main - 34" MSI Mag341CQ (UWQHD) | Second - 24" Asus VX24A
Board Power Draw:- ~30W
View attachment 401859

Happy to finally see more 5000 series Palit numbers, those have been in deficit.
 
MSI RTX 5070ti Ventus 3X OC
Monitor 2560x1440 at 120 hz

It idles at 22 Watt (Windows 11 24H2, driver 576.52, monitor 2560x1440 at 120 Hz)
 
5080 Palit GamingPro
1440p@240hz + 1440p@165hz x2
1.jpg
 
Back
Top