NVIDIA announced its new RTX 50-series graphics cards at CES 2025 on January 7, with prices ranging from $549 to $2,000. The flagship model, the RTX 5090, is set to deliver significant performance improvements over previous models.
NVIDIA RTX 50-series specifications
The new RTX 5090 Founders Edition will feature 32GB of GDDR7 RAM and 21,760 CUDA cores. NVIDIA claims that depending on the game, the 5090 can deliver up to twice the performance compared to the RTX 4090. Demonstrations showed Cyberpunk 2077 running at 238 frames per second on the 5090 with DLSS 4, versus 106 fps on the RTX 4090 using DLSS 3.5.
The RTX 5080 will include 16GB of GDDR7 memory and 10,752 CUDA cores, boasting a total graphics power of 360 watts. NVIDIA positions it as twice as fast as the RTX 4080. The RTX 5070 Ti comes with 16GB of GDDR7 memory and 8,960 CUDA cores, while the RTX 5070 features 12GB of GDDR7 memory and 6,144 CUDA cores. NVIDIA states the 5070 will match the performance of the RTX 4090 due to DLSS 4 capabilities, while the 5070 Ti will be twice as fast as the RTX 4070 Ti.
All RTX 50-series GPUs feature a two-slot design, enhanced cooling with dual flow-through fans, and are equipped with PCIe Gen 5 and DisplayPort 2.1b connectors. This design allows for support of displays up to 8K and 165Hz.
Power requirements and DLSS 4 technology
The RTX 5090 has a total graphics power requirement of 575 watts, with NVIDIA recommending a 1,000-watt power supply unit (PSU). The other models, such as the RTX 5080 and RTX 5070 Ti, have lower total graphics power requirements and PSU recommendations, making them more accessible for a wider range of PCs.
DLSS 4, which NVIDIA highlighted as a key feature of the new series, utilizes AI to boost gaming frame rates. This generation allows for enhanced performance in games that support NVIDIA’s new multi-frame generation technology, while older RTX GPUs will still support DLSS 4, though with limited capabilities compared to the new 50-series cards.
Market context and pricing
The RTX 5090 is priced at $1,999, the RTX 5080 at $999, the RTX 5070 Ti at $749, and the RTX 5070 at $549. NVIDIA indicated a shift in its focus as gaming sales now account for under 10% of total revenue, while data center revenues continue to grow significantly.
NVIDIA’s new gaming chips are designed to incorporate advancements from its data center AI chips, which have seen massive demand across the industry. This integration aims to improve both gaming graphics and performance. The laptop variants of these GPUs will start shipping in March 2025, with prices beginning at $1,299 for devices equipped with the new 5070.
Jensen Huang, NVIDIA’s CEO, emphasized the blending of gaming and AI technologies, highlighting the transformative potential of these new GPUs for both developers and consumers. He stated, “While we are now an AI company as well as a gaming company, our gaming side still benefits tremendously from the fact that we are an AI company.”
Featured image credit: Nvidia
