Previously, Nvidia only had the RTX 6000 Ada Generation workstation graphics solution, which is a fully enabled AD102 GPU with 18,176 CUDA cores and 48GB of ECC GDDR6 memory on a 384-bit memory interface. The new RTX 4000 SFF Ada Generation one is a cut-down version, made to fit a different form-factor, a low-profile dual-slot one with 70W total board power (TBP). Nevertheless, it still provides up to 19.2 TFLOPs of single-precision and 306.8 TFLOPs of Tensor-based compute performance.
In addition to the RTX 4000 SFF Ada Generation, Nvidia also announced a whole range of mobile workstation GPUs, including the RTX 5000, RTX 4000, RTX 3500, RTX 3000 and the RTX 2000 Ada Generation. These are pretty much in line with Ada generation consumer-oriented counterparts, ranging from 35W to 175W TDP, all with ECC memory, excluding the RTX 2000 Ada Generation. The only other exception is the RTX 3500, with 5,120 CUDA cores, which does not have a consumer/desktop counterpart.
According to Nvidia, the RTX 4000 SFF Ada Generation should be available to partners soon, priced at $1250, while laptops with mobile RTX Ada Generation GPUs should also start to appear soon.