Facebook
Reddit
Email
Discover everything you need to know about the newly announced Nvidia RTX 50 Series GPUs at CES 2025. Explore specs, pricing and more.
Nvidia showed off its highly anticipated RTX 50 Series GPUs at CES 2025, and tech enthusiasts and potential buyers gathered to see what was in store. This new series raises the bar with game-changing performance and groundbreaking technology to take gaming and creative workflows to places never seen before.
If you want to get a better understanding of what to expect from the latest additions to Nvidia’s graphics card lineup, here’s a complete overview. In this post, we shall go over the detailed specifications as well as pricing information on each GPU unveiled at CES 2025.
“Blackwell, the engine of AI, has arrived for PC gamers, developers and creatives,” said Jensen Huang, founder and CEO of NVIDIA. “Fusing AI-driven neural rendering and ray tracing, Blackwell is the most significant computer graphics innovation since we introduced programmable shading 25 years ago.”
Nvidia RTX 5090
The GeForce RTX 5090 GPU is the fastest GeForce RTX GPU available with groundbreaking advancements in graphics processing technology. It has an awesome 92 billion transistors in its design and delivers a mind-blowing computing power of over 3,352 trillion AI operations per second (TOPS).
The Blackwell architecture is this innovative architecture that increases processing efficiency and speed, making this a remarkable performance. DLSS 4 (Deep Learning Super Sampling) also introduces, adding up to twice the performance of the GeForce RTX 4090 GPU with the GeForce RTX 5090 GPU. In this leap in technology, gaming experiences get better, and so too does the ability for professional graphics rendering and AI-driven applications.
Specifications
Below are the detailed specifications of the Nvidia RTX 5090 GPU that were shared by Nvidia Founder and CEO, Jensen Huang, in his keynote speech at CES 2025.
| Specs | Details |
| NVIDIA Architecture | Blackwell |
| DLSS | DLSS 4 |
| AI TOPS | 3352 |
| Tensor Cores | 5th Gen |
| Ray Tracing Cores | 4th Gen |
| CUDA Core Count | 21760 |
| Memory Configuration | 32 GB GDDR7 |
| Memory Speed | 28 Gbps |
| Memory Bandwidth | 1792 GB/s |
| TGP | 575W |
| Power Interface | 11 12V-2×6 (16-Pin) |
Pricing
This year’s flagship GPU by Nvidia, the RTX 5090, has been announced at the price tag of $1,999. This is certainly a hefty price to pay, but it is to be expected from a top-of-the-line GPU that is capable of 2x the performance output of the RTX 4090, Nvidia’s previous flagship.
The RTX 5090 and RTX 5080 will be available to be purchased starting January 30th, 2025. The Founders edition can be directly purchased from Nvidia as well as select retailers worldwide.
Nvidia RTX 5080
Designed to be twice as fast as the RTX 4080, the RTX 5080 packs all the latest specifications, including 16GB of its advanced GDDR7 memory and a memory bandwidth of 960GB per second. It has 10,752 CUDA cores and shines when having to handle complex graphics as well as computational tasks.
Nvidia rates the total graphics power at 360 watts and recommends a power supply of at least 850 watts. These will offer great performance gains over the previous RTX 4080 model, bringing a new level of gaming and professional graphics.
Specifications
Following are the specifications that were announced on the Nvidia RTX 5080 GPU, announced by Jensen Huang, Nvidia Founder and CEO.
| Specs | Details |
| NVIDIA Architecture | Blackwell |
| DLSS | DLSS 4 |
| AI TOPS | 1801 |
| Tensor Cores | 5th Gen |
| Ray Tracing Cores | 4th Gen |
| CUDA Core Count | 10752 |
| Memory Configuration | 16 GB GDDR7 |
| Memory Speed | 32 Gbps |
| Memory Bandwidth | 960 GB/s |
| TDP | 360W |
| Power Interface | 1 12V-2×6 (16-Pin) |
Pricing
The Nvidia RTX 5080 boasts a competitive price point of $999. At this budget, users will be able to enjoy about twice the performance of the RTX 4080. This is a significant leap in performance while keeping the cost of the GPU less than predicted before launch. Though the announced prices would make enthusiasts and potential buyers happy, it is yet to be seen how the retail prices fare against them.
Nvidia RTX 5070 Ti
The RTX 5070 Ti packs 16GB of GDDR7 memory with an 896GB/s memory bandwidth. It also boasts 8,960 CUDA cores as well, making it ideal for games and 3D rendering. The RTX 5070 Ti requires a system utilizing a 750-watt power supply unit (PSU). The GPU demands a total graphics power requirement of 300 watts. Nvidia says the RTX 5070 Ti is 2x faster than the RTX 4070 Ti, emphasizing speed and efficiency.
Specifications
Nvidia announced the specifications of the brand-new RTX 5070 Ti at CES 2025.
| Specs | Details |
| NVIDIA Architecture | Blackwell |
| DLSS | DLSS 4 |
| AI TOPS | 1406 |
| Tensor Cores | 5th Gen |
| Ray Tracing Cores | 4th Gen |
| CUDA Core Count | 8960 |
| Memory Configuration | 16 GB GDDR7 |
| Memory Speed | 28 Gbps |
| Memory Bandwidth | 896 GB/s |
| TDP | 300W |
| Power Interface | 1 12V-2×6 (16-Pin) |
Pricing
With the Nvidia RTX 5070 Ti, the graphics processing power, efficiency, and time taken of the GPU surpass those of the RTX 4090. The RTX 5070 Ti is priced at $749. This is definitely revolutionary in terms of performance per buck that you will get in return for an extremely affordable price point for a flagship Nvidia RTX Blackwell Family GPU. The RTX 5070 Ti and RTX 5070 will be available for purchase starting in February. The Founders Edition is to be purchased directly from Nvidia and select retailers worldwide.
Nvidia RTX 5070
With 12 GB of GDDR7 memory and a memory bandwidth of 672 GB/s the RTX 5070 graphics card is great for high-resolution textures and demanding applications. It has 6,144 CUDA cores, which means it’s great at parallel processing for gaming, video editing, 3D rendering, etc.
Its power consumption is 250 watts and will need a PSU of at least 650 watts to run stable. According to NVIDIA CEO Jensen Huang, the RTX 5070 will be able to match the performance of the RTX 4070. No doubt this performance leap is all thanks to DLSS 4 technology. This technology increases frame rates and visual quality without relying on traditional rasterization alone.
Specifications
Following are the hardware specifications of the RTX 5070 GPU by Nvidia, which made its debut at CES 2025.
| Specs | Details |
| NVIDIA Architecture | Blackwell |
| DLSS | DLSS 4 |
| AI TOPS | 988 |
| Tensor Cores | 5th Gen |
| Ray Tracing Cores | 4th Gen |
| CUDA Core Count | 6144 (When enabled fully) |
| Memory Configuration | 12 GB GDDR7 |
| Memory Speed | 28 Gbps |
| Memory Bandwidth | 672 GB/s |
| TDP | 250W |
| Power Interface | 1 12VHPWR (16-Pin) |
Pricing
The Nvidia RTX 5070 is one of the most value-for-money GPUs we have come across in GPU history. This is thanks to AI increasing performance and power efficiency. The RTX 5070 comes at an affordable price of $549. The fact that its performance is comparable to the RTX 4090, Nvidia’s previous flagship GPU, priced at $1,499 at launch, is certainly miraculous. It opens up a lot of avenues for gamers, developers, and creatives to buy an extremely capable GPU at a mere $549.
Looking For More?
Thank you for reading the article. We provide the latest news and create guides for Baldur’s Gate 3, Starfield, ARK Survival Ascended, and more. Also, watch Deltia play games on Twitch or visit his YouTube channel!
