NVIDIA has finally unveiled its long-awaited update to the workstation series, introducing three fresh models to the lineup. Leading the charge is the RTX 5000, housing the ADA AD102 GPU with an impressive 12,800 CUDA cores. Just below the flagship RTX 6000, the RTX 5000 boasts 32GB of memory and is set to retail around $4000, while the RTX 6000 offers 48GB memory.
Expanding the RTX 4000 ADA series, NVIDIA has introduced two new models: the RTX A4500 and the revamped RTX 4000. The latter, previously available in a Small Form Factor (SFF) design, now features a single-slot design departing from its predecessor’s low-profile build. Despite a power requirement increase from 70W to 130W, the boost clock has significantly risen to 2.2 GHz from the previous 1.56 GHz on the SFF variant.
The RTX 4500 ADA comes equipped with the AD104 GPU featuring 7680 CUDA cores and paired with 24GB of memory. Notably, this model integrates a 192-bit memory bus to accommodate the 24GB memory capacity, while power consumption is reported to be around 210W.
Regarding the NVIDIA RTX 4000 & 5000 ADA models, it’s important to highlight that all of these cards will support a complete PCIe Gen4 x16 interface and will come with a quad DisplayPort 1.4a setup. The only distinction is that the RTX 4000 incorporates a mini-DisplayPort, whereas the other models use the standard full-sized connector. The pricing of these cards is aligned with their specific market segments. Following the launch of the $1250 RTX 4000 SFF and the $6800 RTX 6000, the RTX 4500 and RTX 5000 will be priced at $2250 and $4000, respectively. The RTX 4000 non-SFF model will maintain the same pricing as its compact counterpart.
The initial batch of systems featuring the new RTX ADA models is slated for shipment in the upcoming autumn season. Professionals can look forward to these systems being offered by manufacturers like Boxx, Dell, Lambda, Lenovo, and HP. Notably, NVIDIA suggests the possibility of systems even housing up to four RTX 6000 ADA GPUs, boasting a substantial 192GB of VRAM, ideal for handling advanced workloads.
Via NVIDIA