NVIDIA GeForce RTX 3070 Ti Founders Edition video card front side

Introduction

It’s “Ti”ME for the RTX 3070 to get “TIE’d.”  At Computex 2021 NVIDIA announced both the GeForce RTX 3080 Ti and the GeForce RTX 3070 Ti.  On June 2nd we reviewed the GeForce RTX 3080 Ti Founders Edition with an MSRP of $1,199.  Our review compared the video card with the GeForce RTX 3080, GeForce RTX 3090, and AMD Radeon RX 6900 XT. 

Today we have the new NVIDIA GeForce RTX 3070 Ti Founders Edition which is launching today.  Partner cards should be available the day after.  The NVIDIA GeForce RTX 3070 Ti Founders Edition is landing with an MSRP of $599.  For comparison, the GeForce RTX 3070 Founders Edition has an MSRP of $499, and the GeForce RTX 3080 Founders Edition has an MSRP of $699.  This slots into a price point that was previously unoccupied by NVIDIA. 

Compared to AMD video cards, the AMD Radeon RX 6800 has an MSRP of $579 and is, therefore, the closest competitor from AMD.  The RX 6800 XT has an MSRP of $649.  Therefore, it is only $50 more (by MSRP) compared to the new GeForce RTX 3070 Ti.  Keep in mind on MSRP’s, these are simply manufacturer-suggested retail prices and don’t reflect current street pricing.

NVIDIA GeForce RTX 3070 Ti Specifications

The GeForce RTX 3070 Ti is the GeForce RTX 3070 taken to the max.  It beefs it up quite a bit.  It is based on GA104, the same as the RTX 3070.  However, it beefs up practically everything about it, including the overall design.  It has a larger cooler layout than the GeForce RTX 3070.  It’s a bigger video card and has a more robust cooling apparatus to match.  It’s not as big as the RTX 3080, but it is bigger than the RTX 3070.

NVIDIA GeForce RTX 3070 Ti Founders Edition specification presentation slide

The GeForce RTX 3070 Ti is based on the Samsung 8nm GA104-400 die.  It has 6,144 CUDA Cores, 96 ROPs, 192 TMUs, 48 RT Cores (2nd Gen), and 192 Tensor Cores (3rd Gen.)  It operates at a Boost Clock of 1770MHz.  It has 8GB of GDDR6X memory at 19GHz on a 256-bit memory bus providing 608GB/s of bandwidth.  The TDP is 290W. 

GeForce RTX 3080 FE GeForce RTX 3070 Ti FE GeForce RTX 3070 FE
Process/Die/Arch. 8nm GA102-200 Ampere8nm GA104-400 Ampere8nm GA104-300 Ampere
CUDA Cores870461445888
ROPs969696
TMUs272192184
RT Cores684846
Tensor Cores272192184
Boost Clock 1710MHz 1770MHz1725MHz
VRAM 10GB GDDR6X 19GHz8GB GDDR6X 19GHz8GB GDDR6 14GHz
Bus and Bandwidth 320-bit 760GB/s 256-bit 608GB/s256-bit 448GB/s
TDP 320W 290W220W
NVIDIA GeForce RTX 3070 Ti Founders Edition Specification Comparison Table

For comparison to the GeForce RTX 3070, the new GeForce RTX 3070 Ti increases the CUDA Core count to 6144 from 5888 and increases the TMUs to 192 versus 184.  The RT Cores get a bump up to 48 from 46 and the Tensor Cores get a bump up to 192 from 184.  The Boost Clock also gets a slight boost up to 1770MHz from 1725MHz.  Though the real-world frequency will be different.  ROPs stay the same at 96. 

The memory capacity stays the same at 8GB but it is replaced with the newer GDDR6X modules and boosted up to 19GHz from the GDDR6 at 14GHz on the GeForce RTX 3070.  This increases the memory bandwidth to 608GB/s versus 448GB/s on RTX 3070.  The overall TDP also increases from 220W to 290W.  GDDR6X memory consumes more power than GDDR6.

GeForce RTX 3070 Ti Founders Edition Pictures

The looks of the GeForce RTX 3070 Ti Founders Edition are deceiving at first glance. It has the same pattern of styling as the GeForce RTX 3070 Founders Edition, but with the cooler configuration of the GeForce RTX 3080 Founders Edition, but a bit smaller. The GeForce RTX 3070 Founders Edition has a cooler configuration as dual fans on the front of the video card. However, the new GeForce RTX 3070 Ti Founders Edition uses the flipped fan design from the GeForce RTX 3080 Founders Edition and up. That is it has one fan on the front, and one on the back pulling air through. Yet, at the same time, the pattern for the cooler matches the GeForce RTX 3070 Founders Edition. You’ll see some side-by-side pictures below.

This is the PCB of the GeForce RTX 3070 Ti Founders Edition. You can see that it is shorter than the heatsink. It has the unique cut-out at the end to allow the fan to pull in air through the heatpipes to the other side. This PCB design is not the same as the GeForce RTX 3080 Founders Edition however, it’s a different size.

The GeForce RTX 3070 Ti Founders Edition has three DisplayPorts 1.4a and one HDMI 2.1 port. It does use NVIDIA’s 12-pin power adapter. This adapter requires two 8-pin PCI-Express Power Connectors to operate. The NVIDIA 12-pin adapter is not different from what you get on all the Founders Edition cards, therefore any adapter from them can be plugged into this erroneously. For example, you could, by accident, play an adapter from the GeForce RTX 3070 into this video card. However, it only provides 1 8-pin connector, and wouldn’t work. However, you could still physically plug it in because the 12-pin connector part is the same on all the video cards. This design does have that flaw, since you could plug a connector in that wouldn’t work. Make sure you use the dual-8pin one.

GeForce RTX 3080 FE GeForce RTX 3070 Ti FE GeForce RTX 3070 FE
Length 11.2 inches10.5 inches9.5 inches
Width4.4 inches4.4 inches4.4 inches
Height2-slot2-slot2-slot
NVIDIA GeForce RTX 3070 Ti Founders Edition Size Comparison Table

As you can see, the new GeForce RTX 3070 Ti Founders Edition is longer than the RTX 3070 by an inch, but still shorter than the GeForce RTX 3080 Founders Edition. This is because it uses a cooler configuration similar to the GeForce RTX 3080, so it is more beefed up, but it shares the same frame design as the RTX 3070. Where the RTX 3080 has a big “X” metal frame design, the RTX 3070 and RTX 3070 Ti have more of an “S” design.

Brent Justice

Brent Justice has been reviewing computer components for 20+ years, educated in the art and method of the computer hardware review he brings experience, knowledge, and hands-on testing with a gamer oriented...

Join the Conversation

18 Comments

  1. Micron really needs to start making/supplying 2GB GDDR6X chips.

    Still massively disappointed by the efficiency. I’m assuming these are all the 3080 mobile chips that didn’t bin well.

    For 20% more money, it should be reasonably better and just isn’t.

  2. Micron really needs to start making/supplying 2GB GDDR6X chips.

    Still massively disappointed by the efficiency. I’m assuming these are all the 3080 mobile chips that didn’t bin well.

    For 20% more money, it should be reasonably better and just isn’t.

    GA 104 chip. Not the GA 102 that the 3080 uses.

  3. Yar, not sure if I can get excited for any of these GPU launches. Most excited I’ve been recently for a GPU was having my queue come up for a EVGA 3080 Hybrid at MSRP (which promptly replaced my 3090).
  4. Mobile variants are typically not related to their desktop parts (to my knowledge). Power constraints and all.

    Sometimes they are – just with a different TDP or thermal caps. With all the boost algorithms in play these days, you could drop a full 3090 in a laptop and it would run — just your performance would be dependent on the cooling solution and power delivery.

  5. I would be away when all these things come out… I want to get back home to my computer and my local microcenter!!! Great review Brent, interesting card as well. Hopefully availability will be better… fingers crossed.
  6. I don’t think I’d consider this card even if I could get it at retail price.

    My hope is to get a EVGA variant and then step it up to a 3080 ti. At this point msrp and availability of anything is basically a myth so I’ll take one if I can get one, as it will be an upgrade of my 6700XT that I massively overpaid for as well.

  7. My hope is to get a EVGA variant and then step it up to a 3080 ti. At this point msrp and availability of anything is basically a myth so I’ll take one if I can get one, as it will be an upgrade of my 6700XT that I massively overpaid for as well.

    Did you que up today?

  8. Did you que up today?

    I signed up for the new egg shuffle. I’m not evga elite and too lazy to go make like 1000 posts on their forums to get it for free. Guess I should have registered my 1660 Super or 980 when I bought them…

  9. Mobile variants are typically not related to their desktop parts (to my knowledge). Power constraints and all.

    Yep. Pascal was a notable exception. It was a big deal at the time that the mobile chips were the same as the desktop.

  10. Good review and conclusions. 8gb for a $600 card, not street value really does indeed seem limiting. Seems like the best real bang per buck if one can get one is either the 6800 XT (only $49 more) or the 3080 FE, $100 more but way way more performance and at least 10gb of vram.

    Now I do think Nvidia is indeed in a rock and hard spot, the card does use more expensive ram, higher binned die with a little bit more cuda cores, better cooler plus at a time where supply constraints, weaker dollar, inflation, shipping costs (sky rocketed). One really just needs to decide if today is a good day for the price for these cards. Waiting maybe not a good option either as inflation appears to be occurring and at a faster rate than expected. I would recommend the 6800 but more so the 6800 XT over this as well as a 3080 (If all were at MSRP which is a tall order to get).

  11. I have been hearing rumours the FE card is pretty rare, to the point that there are claims Nvidia only launched enough to be able to say they launched them but they leave the real launch to the board partners so those can make a quick cash grab.

    This seems to be mostly related to the fact that there are almost no cards to be found on Ebay and such, while the other cards are still plentifull there.

Leave a comment