# Nvidia Relies Significantly on DLSS 4 and AI-Driven Frames for Performance Comparisons
Nvidia has once more advanced the frontiers of GPU technology with the debut of its RTX 50-series, built upon the new Blackwell architecture. The announcement, which encompasses four new desktop GPUs along with their mobile equivalents, showcases the company’s ongoing commitment to AI-enhanced features, notably the launch of **DLSS 4** and its Multi Frame Generation capability. While Nvidia claims remarkable performance improvements, the dependency on AI-generated frames for performance comparisons has ignited debates regarding the actual real-world implications of these innovations.
## Pricing and Lineup Summary
The RTX 50-series GPUs deliver both advantages and disadvantages for gamers and PC enthusiasts. On the brighter side, three out of the four new GPUs are priced the same as or slightly below their RTX 40-series counterparts:
– **RTX 5080**: $999 (identical to the RTX 4080 Super)
– **RTX 5070 Ti**: $749 ($50 less than the RTX 4070 Ti Super)
– **RTX 5070**: $549 ($50 lower than the RTX 4070 Super)
Nonetheless, Nvidia’s premier **RTX 5090** carries a steep price of $1,999, a considerable jump from the $1,599 MSRP of the RTX 4090. This marks the 5090 as pricier than four top-tier gaming consoles combined or even two midrange gaming computers.
### Core Specifications
The RTX 50-series GPUs feature notable improvements across the entire lineup. Here’s a brief comparison of the flagship RTX 5090 alongside its counterparts:
| Feature | RTX 5090 | RTX 5080 | RTX 5070 Ti | RTX 5070 |
|——————|—————-|—————-|—————-|—————-|
| CUDA Cores | 21,760 | 10,752 | 8,960 | 6,144 |
| Boost Clock | 2,410 MHz | 2,617 MHz | 2,452 MHz | 2,512 MHz |
| Memory Size | 32GB GDDR7 | 16GB GDDR7 | 16GB GDDR7 | 12GB GDDR7 |
| Memory Bus Width | 512-bit | 256-bit | 256-bit | 192-bit |
| Memory Bandwidth | 1,792 GB/s | 960 GB/s | 896 GB/s | 672 GB/s |
| TGP | 575 W | 360 W | 300 W | 250 W |
The RTX 5090 is a powerhouse, housing 92 billion transistors, twice the CUDA cores of the RTX 5080, and a vast 512-bit memory bus. Nvidia has also succeeded in minimizing the physical dimensions of the 5090 Founders Edition, turning it into a dual-slot card that fits more conveniently into compact PC configurations.
## DLSS 4: Transformative Innovation or Marketing Gimmick?
The highlight of the RTX 50-series is **DLSS 4**, which offers **Multi Frame Generation**. This technology enhances the Frame Generation feature from DLSS 3, enabling the GPU to create up to three additional AI-generated frames for every frame rendered traditionally. Nvidia asserts that this allows the RTX 5070 to achieve performance similar to the RTX 4090, a daring claim that is heavily reliant on DLSS 4’s performance.
### Mechanism of DLSS 4
DLSS (Deep Learning Super Sampling) employs AI to upscale images from lower resolutions to higher ones, enhancing performance without compromising visual fidelity. DLSS Frame Generation advances this by producing entirely new frames via AI, thereby effectively boosting the frame rate. With DLSS 4’s Multi Frame Generation, Nvidia states that “15 out of every 16 pixels” appearing on the display are created by AI, harnessing the quicker Tensor Cores of the Blackwell architecture.
While this method can dramatically elevate frame rates, it prompts inquiries about the credibility of Nvidia’s performance assertions. AI-generated frames may occasionally fall short of the quality seen in traditionally rendered frames, and their effectiveness is contingent on game compatibility and the base frame rate.
### Practical Considerations
Nvidia’s dependence on DLSS-based benchmarks has attracted criticism for potentially being misleading. For instance, the assertion that the RTX 5070 matches the performance of the RTX 4090 holds true only in scenarios where DLSS 4 is entirely supported and activated. In the absence of DLSS, the raw performance output of the 5070 is improbable to rival that of the 4090.
## Mobile RTX 50-Series GPUs
Nvidia has also introduced mobile variants of the RTX 50-series GPUs, which are set to be available in laptops ranging from $1,299 to $2,899 starting in March.
