Graphics Cards

NVIDIA GeForce RTX 4060 Ti specs updated to 160W TDP, now less power than 3060 Ti – VideoCardz.com

Please note that this post is tagged as a rumor.

The RTX 4060 Ti could consume less power than predecessor

According to hardware leaker Kopite7kimi, the company has changed the TDP specs for its upcoming mid-range GPU.

The GeForce RTX 4060 Ti is to be the first desktop card to feature an AD106 GPU. The card reportedly features 4352 CUDA cores, and it is most likely to be equipped with 8GB GDDR6 memory. This would also be the first desktop card from this generation not to use GDDR6X technology.

Kopite’s specs from last month have supposedly been updated. For all we know is that the TDP has changed from 220W to just 160W. That’s a significant power reduction for sure, to the point where the successor will consume not 20W more but 40W less power, as it is now claimed.

NVIDIA is now set to launch its first mobile RTX 40 graphics cards on February 8th. One of the card launching on this day is RTX 4070 Laptop GPU, which is reportedly using GN21-X6 GPU also known as AD106. This would suggest that NVIDIA already has the GPU ready, and nothing should stop them from releasing it to the desktop market as well. Nothing except the existing stock of RTX 3060 series.

As long as rumored specs are considered, the RTX 4060 Ti would have fewer cores and half of the memory bus. It is not unclear how would this affect performance, but the bandwidth will be heavily reduced to just 288 GB/s (RTX 3060 Ti has 608 or 448 GB/s depending on SKU).

The RTX 4070 non-Ti and RTX 4060 Ti are both expected in late Q1 2023 or early Q2. These cards will finally bring Ada Lovelace architecture to lower price points, which this architecture desperately needs.

RUMORED NVIDIA GeForce RTX 40 Series Specs
VideoCardz.com RTX 4070 Ti RTX 4070 RTX 4060 Ti RTX 3060 Ti
Picture
Board-SKU PG141-SKU331 PG141-SKU336/337 PG190 PG142
Architecture Ada (TSMC 4N) Ada (TSMC 4N) Ada (TSMC 4N) Ampere (SAMSUNG 8N)
GPU AD104-400 AD104-250/251 AD106-350 GA104-200/202
CUDA Cores
Boost Clock

2.6 GHz

TBC TBC

1.67 GHz

Memory

12 GB G6X

12 GB G6X

 

8 GB G6

8 GB G6 or G6X

Memory Bus

192-bit

192-bit

 

128-bit

256-bit

Memory Bandwidth

504 GB/s

504 GB/s

 

288 GB/s

448 or 608GB/s

Default TGP

 

220W→160W

Release Date January 5th, 2023 Q1 2023 (?) Q1 2023 (?) December 2020

Source: @kopite7kimi