NVIDIA’s highly anticipated GeForce RTX 50 “Blackwell” GPUs might face a shift in their launch timeline due to a focus on data center technology. These next-generation graphics cards are eagerly awaited by gamers and tech enthusiasts looking to upgrade their current setups. Despite the excitement, there could be delays caused by changes in the design, linked to prioritizing data center chip enhancements.
Recent updates indicate that NVIDIA has revised its Data Center Blackwell AI chips, addressing issues associated with the GPU mask. With the redesign completed, NVIDIA is set to begin volume production in Q4 2024, aiming for significant revenue.
Reports from industry insiders via MyDrivers suggest that the redesign might also impact consumer GPUs. The main problem was a thermal expansion coefficients mismatch among various components, leading to chip deformation and system failures. NVIDIA resolved this by redesigning the top metal layer and bump structures.
Despite the distinction between the Data Center and Consumer tier Blackwell chips, they share a similar core architecture. This similarity could mean that the upcoming GeForce RTX 50 series might experience some effects from these changes. However, NVIDIA assures that the samples sent for testing are the corrected versions, and the units heading into mass production feature the updated design.
NVIDIA is currently prioritizing Data Center “Blackwell” GPUs due to their substantial revenue contributions and high demand, with production capacity limitations at TSMC being the only constraint.
Here’s a preliminary look at the potential specs for the upcoming NVIDIA GeForce RTX 50 “Blackwell” GPU lineup:
– **GeForce RTX 5090 Ti?**: Possibly featuring GB202 GPU, 192 SMs/Cores, 512-bit GDDR7 memory configuration, with a TDP of 450W.
– **GeForce RTX 5090?**: Could include GB202 GPU with a 448-bit GDDR7 memory configuration, TDP of 400W.
– **GeForce RTX 5080 Ti?**: Speculated to use GB202 GPU, 384-bit GDDR7 memory configuration, TDP of 350W.
– **GeForce RTX 5080?**: Expected to have GB203 GPU, 96 SMs/Cores, 256-bit GDDR7 memory configuration, TDP of 320W.
– **GeForce RTX 5070?**: Might use GB205 GPU with a 192-bit GDDR7 memory configuration, TDP of 225W.
– **GeForce RTX 5060 Ti?**: Potentially featuring GB206 GPU, 128-bit GDDR6 memory configuration, TDP of 160W.
– **GeForce RTX 5060?**: Likely to include GB207 GPU with a 128-bit GDDR6 memory configuration, TDP of 120W.
Previous mentions indicated that the release of GeForce RTX 50 “Blackwell” GPUs might be pushed to CES 2025. Kopite7kimi, a reliable source of NVIDIA leaks, has also mentioned an increase in power consumption in the new chips, particularly in higher-end models. However, it’s unclear if this increase is only seen in prototype units or will persist in final retail versions.
Regardless, anticipation for the RTX 50 series remains high among the gaming community. Everyone is eager to see what new advancements NVIDIA will bring to the market with these powerful GPUs.
Which NVIDIA GeForce RTX 50 GPU are you most excited about? Share your thoughts on social media!
