NVIDIA is set to introduce two new models in its GeForce RTX lineup: the GeForce RTX 5060 Ti and RTX 5060, both boasting upgraded memory configurations with 9 GB of capacity.
Enhanced Memory for NVIDIA’s GeForce RTX 5060 Ti & RTX 5060 GPUs
Fresh insights suggest that NVIDIA is expanding memory options beyond the previously reported 9 GB variant of the GeForce RTX 5050. New developments have emerged regarding the GeForce RTX 5060 Ti and RTX 5060, which will be enhanced with additional memory.
According to recent reports sourced from Chinese Board Channel forums, NVIDIA is actively preparing to launch the upgraded versions of these GPUs. Even though the original models are currently available in retailers, these forthcoming updates will increase memory capacity.
Reports indicate that as Samsung and Micron ramp up production of 3 GB GDDR7 modules, NVIDIA’s add-in board partners (AIBs) are being instructed to upgrade memory configurations for the 5060 series. This means that both the GeForce RTX 5060 Ti and RTX 5060 will transition from an 8 GB to a 9 GB memory capacity. Notably, while the RTX 5060 Ti does come with a 16 GB variant, it is priced $50 higher than the standard 8 GB model.
NVIDIA’s RTX 50 Series May Adopt a 3GB VRAM Module Strategy
The three major upstream memory manufacturers have already established production capacity for 3GB GDDR7 modules. NVIDIA plans to employ a multi-vendor strategy to alleviate the price pressures caused by VRAM shortages; furthermore, 3GB modules allow for increased memory capacity without expanding the memory bus width, representing an optimal balance between cost and performance.
Compared to traditional 2GB modules, a single 3GB GDDR7 module can increase VRAM capacity by 50% at the same bus width, thereby resolving VRAM bottlenecks in high-resolution gaming and AI inference tasks. As manufacturers such as Samsung and Micron ramp up their production capacities, these modules are expected to debut on the market in the near future; NVIDIA has already officially incorporated this strategy into its product roadmap.
According to internal sources within the upstream supply chain, 3GB GDDR7 modules are currently in the product planning phase. It is anticipated that related new products featuring this technology could officially launch around late May or early June.
Reportedly, the RTX 5090 series for laptops has already begun utilizing 3GB VRAM modules. This time, the adoption will extend directly to other models within the RTX 50 series—specifically, mainstream consumer-grade models such as the RTX 5060 Ti 8GB and RTX 5060 8GB series may directly switch to using 3GB modules. These models could be configured with three 3GB modules each, resulting in a planned total VRAM capacity of 9GB. This approach allows for increased memory capacity while keeping the GPU core and other specifications unchanged, serving as the optimal solution for mitigating the severe price hikes and supply shortages currently affecting VRAM.
Machine Translated via Channel Boards
Evaluating the specifications, the introduction of 9 GB memory comes at a trade-off concerning bandwidth. The current 8 GB models of the RTX 5060 Ti and RTX 5060 utilize a 128-bit bus interface for bandwidth. Transitioning to a 9 GB memory configuration means employing three 3 GB modules across a 96-bit bus interface. If NVIDIA retains the performance speeds established in prior models (28 Gbps), the total bandwidth would reach approximately 336 GB/s.
When comparing this to the existing 128-bit bus models, which provide 448 GB/s of total bandwidth, this translates to a significant 33% reduction in bandwidth—a critical factor as these graphics cards may face performance constraints. Even if NVIDIA increases speeds to 30 Gbps, this would yield a bandwidth of merely 360 GB/s, still falling short of the current models. Consequently, the new GPUs will have to sacrifice some available bandwidth to gain the additional GB of VRAM.
Present reports suggest that NVIDIA’s add-in card vendors will aim to launch the GeForce RTX 5060 Ti 9 GB and RTX 5060 9 GB models around May to June. More concrete information is anticipated to surface during this year’s Computex event.
For more details, refer to the news source: Gazlog.
Leave a Reply