NVIDIA’s next-generation GeForce RTX 50 “Blackwell” Gaming GPU flagship is expected to include GDDR7 memory support and will most likely be expensive
The NVIDIA GeForce RTX 50 series is expected to use the Blackwell Gaming GPU architecture, though it will be slightly different from the one that will be available to HPC and AI customers next year. There have been roadmaps that point to a 2025 launch, but some rumors suggest that we may see the latest graphics lineup in late 2024.
Recent revelations by Kopite7kimi point to the NVIDIA GeForce RTX 50 Gaming GPU Flagship with GDDR7 Memory Support.
Next-Generation Memory for Next-Generation Gaming GPUs!
Today’s topic, on the other hand, is about memory configuration, which, according to Kopite7kimi, will ensure the latest memory standard. The GDDR7 memory standard will be an evolution of the GDDR6 and GDDR6X standards, with faster pin speeds and denser capacities.
The initial dies employ up to 24 GB modules and speeds of up to 32 Gbps. These are the ones that will make their debut in 2024. A faster revision is planned for 2026, but NVIDIA is unlikely to wait that long to use the most advanced memory technologies. Rather, a new or refreshed family would use it.
The bandwidth provided by 32 Gbps pin speeds across multiple bus configurations is as follows:
- 128-bit @ 32 Gbps: 512 GB/s
- 192-bit @ 32 Gbps: 768 GB/s
- 256-bit @ 32 Gbps: 1024 GB/s
- 320-bit @ 32 Gbps: 1280 GB/s
- 384-bit @ 32 Gbps: 1536 GB/s
- 512-bit @ 32 Gbps: 2048 GB/s
Although the exact memory configurations for the NVIDIA GeForce RTX 50 Flagship GPU are not specified, it is stated that the card will have a 384-bit bus interface, implying that 512-bit is no longer on the table or is only intended for a higher-end variant. According to the 384-bit memory bus, we will either get the same 24 GB VRAM or 36 GB depending on the DRAM densities used.
The 384-bit bus interface would offer up to 1.5 TB/s of bandwidth at 32 Gbps speeds, representing a more than 50% increase over the existing RTX 4090 GPUs.
The Founders Edition Design Is Similar To The Leaked “4090 Ti” Prototype.
Kopite7kimi also claims that NVIDIA’s GeForce RTX 50 Flagship GPU will have a Founders Edition design, similar to the previously leaked “GeForce RTX 4090 Ti” GPU. The graphics card has a quad-slot cooling solution with a unique side-mounted PCB design and power cables that are routed around the body via large copper links.
If the same design is used for RTX 50 GPUs, we can expect design updates in the final model versus the prototype solution shown below.
The GB202 could also arrive as early as next year, according to the information. The GB203 GPU is said to have a 256-bit bus interface, whereas the GB204 and GB205 GPUs are said to be mutually exclusive.
Yesterday, RedGamingTech highlighted similar specifications, which have now been verified by Kopite, so it appears that things are really shaping up for NVIDIA’s next-gen lineup, though the company is also preparing its GeForce RTX 40 “SUPER” Refresh, which will debut at CES in January 2024 and be released the same month if everything goes as planned.
We shouldn’t expect anything official from NVIDIA on the RTX 50 GPUs this early, but we can expect more updates and leaks as we get closer to the next-gen gaming family.
NVIDIA GeForce GPU SKUs:
| GENERATION | PASCAL | TURING | AMPERE | ADA LOVELACE | BLACKWELL |
| Process Node | TSMC 16nm | TSMC 12nm | Samsung 8nm | TSMC 5nm | TBD |
| Launch Year | 2016 | 2018 | 2020 | 2022 | 2025 |
| Ultra-Enthusiast SKU | GP102 | TU102 | GA102 | AD102 | GB202 |
| Enthusiast SKU | GP104 | TU104 | GA102 | AD103 | GB203 |
| High-End SKU | GP104 | TU106 | GA104 | AD104 | GB205 |
| Mainstream SKU | GP106 | TU106 | GA106 | AD106 | GB206 |
| Entry-Level SKU | GP107 | TU116/117 | GA107 | AD107 | GB207 |