The GTX 1650 is available in five variants, including the GPU-based TU106, TU116, and TU117, as well as a SUPER model. However, not all of these cards are accessible everywhere, and the same approach may be utilized with the GTX 1630. This is the second time we've heard about the GTX 1630, and from what we can see, it appears that it will be released shortly, either by the end of this month or in June.
Update: NVIDIA's GeForce GTX 1630 will ship on May 31. The GTX 1630 is built on the GTX 1650 and has 512 CUDA cores, 4 GB of GDDR6 memory, and a 64-bit memory interface. This is a decrease from the GTX 1650's 896 CUDA cores and 128-bit memory bus, although the boost frequency is 1800 MHz @ 75 W. This memory arrangement has 96 GB/s of theoretical bandwidth, half of the GDDR6 GTX 1650's. NVIDIA's Computex speech next week may introduce the GeForce GTX 1630.