Nvidia has just launched the GeForce RTX 4070 Ti and its little sister, the RTX 4070 has been confirmed by a blunder from Galax. The novelties will not stop there. The firm intends to continue the deployment of its Ada Lovelace architecture on the mid-range through the GeForce RTX 4060 Ti.
This reference has already been talked about, however new speculations have appeared. They reveal important changes transforming its strengths in the face of the current GeForce RTX 3060 Ti.
GeForce RTX 3060 Ti, what do we know?
This card will be the first reference to use an AD106 GPU. It should be equipped with 4,352 CUDA cores accompanied by 8 GB of memory. Nvidia would have opted for GDDR6 and not GDDR6X, a first for this new generation of graphics solution.
| Speculations | GeForce RTX 4060 Ti | GeForce RTX 3060 Ti |
| GPUs | AD106-350 | GA104-200/202 |
| CUDA Cores | 4352 | 4864 |
| Boost frequency | ? | 1.67GHz |
| Memory | 8GB G6 | 8 GB G6 or G6X |
| Memory bus | 128 bit | 256 bit |
| Memory bandwidth | 288 GB/s | 448 or 608 GB/s |
| Default TGP | 220W→160W | 200W |
| Release date | Q1 2023 (?) | December 2020 |
The latest information from Kopite announce a significant drop in the thermal envelope. The TGP plummets from 220W (first rumors) to just 160W. As a result, the card promises to be less greedy by 40 Watts than its elder instead of “more greedy by +20 Watts”.
Regarding a launch date, the case has not yet been finalized. It is possible that February 8, 2023 will be the big day but there is nothing official at the moment. The indiscretions state that the AD106 is operational so nothing is stopping Nvidia from pressing the green light except for one thing…the current stocks of the GeForce RTX 3060 series.