In the shadow of the AI euphoria and tense DRAM markets, NVIDIA is planning a new attack on the upper GPU class with the RTX 5080 SUPER. But what’s behind the upcoming flagship of the SUPER series and why could the launch be bumpier than expected?
As with the RTX 40 SUPER series, NVIDIA is once again using the SUPER models as a tactical refresh, without any major architectural changes, but with massive VRAM upgrades. The focus is on the RTX 5080 SUPER with a whopping 24 GB GDDR7, a huge leap compared to the 16 GB of the RTX 5080. However, the timing is explosive: DRAM shortages, rising memory prices and an oversaturated market could take the wind out of the sails of the release.
Classification of the upcoming model andtechnical assessment of the circulating data
In the current environment of a tight DRAM market and a product strategy increasingly focused on AI, the RTX 5080 SUPER is a typical refresh within the SUPER series. The architecture remains unchanged, but the model stands out due to the significantly increased memory expansion and higher bandwidths. The technical key points circulating on the net speak of a complete GB203 chip with an unchanged computing unit, but with 24 GB GDDR7 instead of the 16 GB of the RTX 5080. The increase in bandwidth due to faster memory is in an area that primarily favours complex scenarios, but is only visible to a limited extent in many classes of everyday applications.
The timing is difficult, as the supply situation for GDDR and DRAM has remained tight for months. Manufacturers and suppliers are reporting rising purchase prices, creating a high probability that the new top-of-the-range SUPER series model will be launched at a significant premium. At the same time, a market glut of existing RTX 50 models could put pressure on demand, as many users have only recently made the switch.
The technical specifications that have been circulated indicate a minimal change in the computing block, while memory and clock rates have been specifically increased. The power budget is said to be well over four hundred watts and would therefore tighten cooling and PSU dimensioning. A significant increase in performance in raster rendering is hardly to be expected due to the identical number of cores. Performance increases only become plausible where bandwidth or a large amount of memory were limiting factors, such as in 4K games with high texture resolutions or in workloads that load large models or data sets.
The RTX 5080 SUPER is apparently based on the complete GB203 chip with 10,752 CUDA cores. The memory clock is to be increased to 32 Gbps (previously 30 Gbps), the bandwidth would increase to 1024 GB/s via a 256-bit interface. The TBP is expected to rise to over 400 watts, so we could be in for another performance monster with corresponding cooling requirements:
| Specification | RTX 5080 SUPER (rumors) | RTX 5080 |
|---|---|---|
| GPU SKU | GB203-450 | GB203-400 |
| CUDA cores | 10.752 | 10.752 |
| VRAM | 24 GB GDDR7 | 16 GB GDDR7 |
| Memory interface | 256-bit | 256-bit |
| GDDR7 speed | 32 Gbps | 30 Gbps |
| Memory bandwidth | 1024 GB/s | 960 GB/s |
| TBP (total consumption) | >400 W | 360 W |
| Clock frequency | Higher (assumed) | Standard |
A general increase of between one and five percent in conventional games therefore seems realistic. Only with memory-bound applications, i.e. complex ray tracing geometry or AI model calculations, does a 24 GB expansion unfold its full potential. For users who play in 1080p or 1440p, however, the added value can hardly be justified.
Target group consideration, price and expected market launch
The RTX 5080 SUPER is clearly positioned between classic gaming and semi-professional tasks with high memory usage profiles. Users who run local AI models or work in 3D rendering pipelines will benefit noticeably from the memory reserves and higher bandwidth. For pure gamers, there is only an advantage if consistent operation in 4K or a strong focus on ray tracing and high-resolution textures is planned. In all other scenarios, the model remains more of a luxury product with manageable practical benefits.
The price zones of between 1,100 and 1,300 US dollars mentioned online appear to be a conservative lower limit in view of the current GDDR market situation. Several industry reports assume a continuing shortage of GDDR7, which makes production more expensive and limits production volumes. A later market launch in the third quarter seems more plausible than an earlier appearance, as the supply situation could be more stable by then. A presentation at CES is therefore less likely than a presentation at Computex.
Final evaluation
The RTX 5080 SUPER shows the typical pattern of a technically solid but limited innovative refresh. The main advantages lie in the memory and bandwidth, while the computing power only increases slightly. The current market makes a sober classification difficult, as high memory prices weigh on the end price and position the card more in a premium segment. The upgrade may make sense for AI users and creative professionals, while gamers should weigh up the additional price carefully. Overall, the model remains a product with clearly defined strengths, but without broadly noticeable advances for the majority of classic gaming users.
Sources:
TrendForce, DigiTimes, various board channel discussions from China, publicly available NVIDIA roadmaps and own sources at the AIC
27 Antworten
Kommentar
Lade neue Kommentare
Urgestein
Veteran
Urgestein
Veteran
Urgestein
Veteran
Urgestein
Urgestein
Urgestein
Moderator
Mitglied
Urgestein
Mitglied
Mitglied
Veteran
Urgestein
Urgestein
Veteran
Veteran
Alle Kommentare lesen unter igor´sLAB Community →