NVIDIA holds a dominant position in the graphics card market, boasting approximately 90% market share, far ahead of competitors like AMD (around 10%) and Intel (less than 1%). Given its strong market standing, NVIDIA might be considering a strategic shift in response to the exorbitant prices of hardware, including GPUs. This speculation is fueled by the emergence of an engineering sample of a potential RTX 3050 Ti with Ampere architecture and 6 GB of VRAM, which utilizes the GA106 chip, the same as the RTX 3060.
NVIDIA’s history, dating back to the late 1990s, saw rapid user acquisition despite being a newcomer. Within a few years, it surpassed ATI Radeon, an earlier competitor. This dynamic continued after AMD acquired ATI in 2006, with NVIDIA consistently maintaining its lead over Radeon graphics cards to this day.
Leaked RTX 3050 Ti 6GB Features GA106 Chip with 3,328 CUDA Cores as an Engineering Sample
The GTX 1000 series represented NVIDIA’s last generation of relatively affordable graphics cards, with high-end models costing under €800 and an 8GB GTX 1070 available for around €400 in 2016. The RTX 20 series saw a significant price increase, and while the Ampere-based RTX 30 series was considered an improvement with slightly better pricing, it was heavily impacted by cryptocurrency mining. In the current market, where prices remain high, the RTX 3060 is still the most used graphics card on Steam, indicating a substantial market for NVIDIA if they were to launch a product like the leaked RTX 3050 Ti.
This rumored RTX 3050 Ti is specified with 3,328 CUDA Cores, making its specifications remarkably similar to higher-tier cards. It indeed uses the GA106 chip, identical to the RTX 3060. However, the primary concern with this RTX 3050 Ti is its listed 6 GB of VRAM, which is likely insufficient for modern gaming demands. It also features 48 ROPs and 104 TMUs. Interestingly, this GPU has been rumored to be in development since 2021, with its appearance surfacing five years later.
This RTX 3050 Ti Was Never Officially Launched by NVIDIA, Utilizes 2020 Ampere Architecture, and Offers Performance Close to an RTX 3060
Information suggests that the RTX 3050 Ti uses Samsung GDDR6 memory, as indicated by GPU-Z. Its clock speeds are listed at 1,410 MHz base and 1,665 MHz Boost, with memory running at 1,750 MHz on a 192-bit bus, providing a bandwidth of 336 GB/s. A comparison with the RTX 3060 reveals many similarities, including the same 192-bit memory bus, although the RTX 3060 operates at a higher memory frequency of 1,875 MHz, resulting in greater bandwidth. The base clock speeds are close, with the RTX 3060 at 1,445 MHz, but its Boost clock reaches 1,860 MHz and it features 3,584 CUDA Cores, providing a performance advantage along with its 12 GB of VRAM.
This difference is reflected in the benchmark results, where the RTX 3050 Ti scores 7,787 points in Time Spy, approximately 10% lower than the RTX 3060, which typically achieves around 8,600 points. The drivers tested with this RTX 3050 Ti were GeForce 516.61 from June 2022. In the hypothetical scenario of an official NVIDIA launch, this card would occupy an unusual market position. It significantly outperforms the 8 GB RTX 3050, which scores around 5,400 points in Time Spy, yet it possesses less VRAM than even that model.
