**Nvidia Launches the RTX 5090 at CES 2025: A New Benchmark in GPU Excellence Comes with a High Cost**
The eagerly awaited Nvidia GeForce RTX 5090 has now been introduced at CES 2025, meeting the expectations of the tech community—a true marvel of innovation and performance. However, Nvidia’s flagship GPU from its Blackwell generation carries a significant price, establishing a new standard for high-end graphics cards. With a suggested retail price of $1,999 in the U.S. and €2,329 in Europe, the RTX 5090 is not only a technical wonder but also a considerable expense for gamers, creators, and professional users.
### **A New Benchmark in GPU Capability**
The RTX 5090 represents the pinnacle of Nvidia’s latest Blackwell architecture, featuring specifications that are truly revolutionary. In comparison to its predecessor, the RTX 4090, the RTX 5090 showcases a 33% rise in CUDA cores, increasing from 16,384 to 21,760. This is further enhanced by a significant memory upgrade, now incorporating 32GB of GDDR7 with a 512-bit interface, offering an impressive memory bandwidth of 1,792 GB/s. This marks a considerable advancement over the RTX 4090’s 24GB of GDDR6X and 1,008 GB/s bandwidth.
In spite of its substantial power, the RTX 5090 retains a surprisingly compact two-slot form factor, making it more space-efficient than several of its high-performance predecessors. However, this reduced size comes at a cost, as the card’s TDP (Thermal Design Power) has increased to 575 watts, a rise from the RTX 4090’s 450 watts. Consequently, users will require a strong power supply to support the GPU’s energy needs.
### **Highlighted Specifications: RTX 5090 Compared to RTX 4090**
| Feature | Nvidia GeForce RTX 5090 | Nvidia GeForce RTX 4090 |
|————————|——————————————|——————————————|
| **Graphics Chip** | GB202-300, 170 SMs, 21,760 CUDA cores | AD102-300, 128 SMs, 16,384 CUDA cores |
| **Clock Speeds** | Base: 2.01 GHz, Boost: 2.41 GHz | Base: 2.23 GHz, Boost: 2.52 GHz |
| **Memory** | 32GB GDDR7, 512-bit, 28 Gbps | 24GB GDDR6X, 384-bit, 21 Gbps |
| **Memory Bandwidth** | 1,792 GB/s | 1,008 GB/s |
| **TDP** | 575W | 450W |
| **Display Outputs** | 3x DisplayPort 2.1a, 1x HDMI 2.1 | 3x DisplayPort 1.4a, 1x HDMI 2.1 |
### **Performance: A Major Advancement with DLSS 4**
Nvidia asserts that the RTX 5090 can achieve up to 200% performance enhancements over the RTX 4090 in specific cases, primarily due to the new DLSS 4 (Deep Learning Super Sampling) feature. DLSS 4 utilizes advanced AI algorithms to elevate lower-resolution visuals in real-time, yielding a notable increase in frame rates without sacrificing image quality. Nonetheless, it’s important to note that these performance improvements are most notable in games that adopt path tracing techniques, such as *Cyberpunk 2077*, *Alan Wake 2*, and *Black Myth: Wukong*.
In more conventional rasterized games like *Far Cry 6* and *A Plague Tale: Requiem*, the performance gains are more moderate, closely mirroring the 33% rise in CUDA cores. This indicates that while the RTX 5090 excels in cutting-edge gaming and rendering applications, its full capabilities will only be unlocked in software that is tailored for Nvidia’s most recent technologies.
### **An Expensive Consideration**
The RTX 5090’s price of $1,999 represents a substantial jump from the RTX 4090’s initial price of $1,599 in 2022. In Europe, the price increase is even more significant, with the RTX 5090 launching at €2,329 compared to the RTX 4090’s €1,949. This firmly places the RTX 5090 in the ultra-premium category, aimed at enthusiasts and professionals who seek top-tier performance.
Interestingly, Nvidia has opted for a different pricing strategy with the RTX 5080, which is said to be less expensive than its predecessor, the RTX 4080. This could position the RTX 5080 as a more appealing choice for gamers in search of high-level performance.