It appears that NVIDIA’s GeForce RTX 40-series will experience a quick ramp-up to high-volume market sectors in the spring and summer. The GeForce RTX 4070, which should dominate the performance sector but is expected to cost an eye-watering $750, will be released by the business in April. Before the 2023 Computex, NVIDIA may release the GeForce RTX 4060 Ti and the RTX 4060 in May of that year.

The “AD106” silicon, which is anticipated to contain about 36 streaming multiprocessors (4,608 CUDA cores, 144 Tensor cores, 36 RT cores, 144 TMUs, and 48 ROPs); and a 128-bit wide memory interface, is the foundation for both the RTX 4060 Ti and RTX 4060. In contrast to the more exotic 19 Gbps GDDR6X, NVIDIA is anticipated to use 18 Gbps-rated standard GDDR6 memory chips with these SKUs, which should result in cheaper prices (for the firm).

32 out of 36 SM are anticipated to be present in the faster GeForce RTX 4060 Ti, translating to 4,352 CUDA cores, 128 Tensor cores, 32 RT cores, 128 TMUs, and 48 ROPs. The less expensive RTX 4060 (non-Ti) is anticipated to be even more condensed, or possibly even an overclocked version of the “AD107” silicon with 3,072 CUDA cores. With 18 Gbps GDDR6 memory, which should deliver 288 GB/s of memory bandwidth through the 128-bit interface, both SKUs are anticipated to come with 8 GB of standard memory. Since the RTX 4060-series is anticipated to have a typical graphics power (TGP) considerably below 200 W, graphics cards may be more frequently seen using conventional 8-pin PCIe power connectors rather than the 16-pin 12VHPWR.