Nvidia's new Pro GPU costs less than $650 — RTX 2000 Ada Generation arrives with 2,816 CUDA cores and 16GB ECC VRAM

Nvidia has expanded its reach in the professional market with a new entry-level workstation graphics card. Debuting today at $625, the RTX 2000 Ada Generation targets professional consumers who own compact systems and are also on a very tight budget.

The RTX 2000 Ada Generation is a small form factor (SFF) graphics card, sharing an identical design and dimensions as the faster RTX 4000 SFF Ada Generation and the previous RTX A2000. The RTX 2000 Ada Generation features a dual-slot design with a blower-type cooling system that measures 6.6 inches (16.76 cm) long. Nvidia includes a standard ATX and low-profile bracket, so the RTX 2000 Ada Generation will fit equally into standard and SFF systems.

Nvidia's choice of interface on the RTX 2000 Ada Generation is interesting, though. The RTX 2000 Ada Generation features a PCIe 4.0 x8 interface. In comparison, the RTX A2000 and the RTX 4000 SFF Ada Generation tap the PCIe 4.0 x16 interface.

Nvidia uses the AD107 silicon inside the RTX 2000 Ada Generation. The AD107 die is home to 24 Streaming Multiprocessors (SMs), equivalent to 3,072 CUDA cores. However, the RTX 2000 Ada Generation's die only has 22 of the 24 SMs enabled, putting the CUDA core count to 2,816. As a result, the RTX 2000 Ada Generation's performance should lie in between a GeForce RTX 4050 Mobile (2,560 CUDA cores) and a GeForce RTX 4060 (3,072 CUDA cores).

RTX 2000 Ada Generation Specifications

The RTX 2000 Ada Generation has around 15% fewer CUDA cores than the RTX A2000 12GB; however, the improved performance is due to the new Ada Lovelace architecture rather than a higher CUDA core count. If we look at the single-precision numbers, the RTX 2000 Ada Generation outperforms the RTX A2000 12GB by a 50% margin.

On paper, we're witnessing higher RT performance on the RTX 2000 Ada Generation with the 3rd Generation RT cores. Nonetheless, the Tensor performance has increased substantially, delivering over 3X on paper thanks to the new 4th Generation Tensor cores.

Nvidia's touting performance uplifts between 1.3X to 1.6X, with the RTX 2000 Ada Generation having the most significant improvements in the generative AI workloads. Compared to the older Quadro P2200, the RTX 2000 Ada Generation delivers 2X higher performance in SOLIDWORKS SPECviewperf 2020 and up to 4X in SOLIDWORKS Visualize benchmarks.

RTX 2000 Ada Generation Benchmarks

While the RTX A2000 was available in 6GB and 12GB variants, the RTX 2000 Ada Generation only comes in the 16GB flavor. While there is more memory on the newer model, the memory bandwidth took a performance hit. The RTX 2000 Ada Generation has a 128-bit interface, a significant downgrade from the 192-bit interface on the RTX A2000. As a result, the RTX 2000 Ada Generation can only supply a memory bandwidth of up to 224 GB/s, which is around 22% lower than the RTX A2000.

The max power consumption hasn't changed between the two generations. The RTX 2000 Ada Generation sticks to 70W, like the RTX A2000 and RTX 4000 SFF Ada Generation, so it doesn't need external power connectors. The outputs are also the same, offering consumers four mini DisplayPort 1.4a outputs.

The RTX 2000 Ada Generation has a $625 MSRP, 22% more expensive than the RTX A2000 when it debuted three years ago with a $449 MSRP. Despite its age, the latter's value hasn't decreased over the years but quite the opposite. New RTX A2000 models are selling for over $632.98, so the RTX 2000 Ada Generation isn't a bad deal if you can find the graphics card at its MSRP.