The increase in the price of central processing units and graphics processing units has resulted in a drop in the cost of the highly-anticipated NVIDIA RTX 4080 graphics card. As the use of TSM’s 3nm node becomes more widespread in the production of CPUs and GPUs, it is possible that there may be price changes.
TSMC has previously stated that the shift to a 3nm node may result in a logic density of 1.6x to 1.7x, which could potentially impact pricing. It is important to consider all available information before making any decisions.
It is worth considering the relatively small decrease in bitsole size for SRAM on the N3 node, which only saw a 5% reduction, when analyzing the potential effects of the N3 and N3E nodes on chip pricing. This is particularly significant as SRAM is used in the production of chip cache, such as L2 and L3, and the bitsole size remained unchanged on the N3E node.
Save on the RTX 4080 Before More Expensive CPUs and GPUs Hit the Market
As processors become more advanced, they often require larger cache sizes, which can be a challenge. In order to reduce the cost of newer nodes, chip manufacturers have historically relied on increasing density to fit all components into a smaller form factor. Unfortunately, the constant size of SRAM could be a significant issue. It is important to take this into consideration when assessing the potential impact of newer nodes on chip pricing.
The challenges associated with scaling SRAM at smaller nodes, as demonstrated by Intel’s upcoming 7nm EUV process, suggest the need for alternative approaches. One such method is the use of multi-chip modules and alternative process nodes for SRAM, a technique that has been utilized by AMD with success.
Despite the adoption of strategies such as multi-chip modules and alternative process nodes, there is still the potential for chipmakers to eventually move to smaller node sizes. This could result in significant price increases for future CPUs and GPUs. While it is hoped that a solution will be found, the current outlook is uncertain.
The specifications of the upcoming NVIDIA RTX 4060Ti GPU, which have been leaked, indicate that it may have a short-reference board and potentially be smaller in size than previous models. This could make it a good option for those looking to build smaller ITX systems.
The NVIDIA RTX 4060Ti GPU is expected to be made using the NVIDIA 8106 GPU (AD106-350-A1) with 4352FP32. It is likely to come with a 16-pin connector, similar to those found on higher-end cards, but with a lower 220W TDP. NVIDIA may be attempting to make this a standard connector for all GPUs. The card is expected to have 32M L2 cache and 8G 18Gbps GDDR6 memory, making it a highly capable option. There have also been benchmarks released for the upcoming CPUs.
The NVIDIA RTX 4070Ti has been rumored to potentially be a rebranded version of the 12GB RTX 4080, as indicated by the recent leak of the PNY 470TI, which appears to be similar to the 12GB RTX 4080. The price of the RTX 4080 has recently decreased again after its initial MSRP was reduced.
