Excitement is building among PC enthusiasts and gamers following a significant leak regarding Nvidia’s anticipated RTX 50 series “super” refresh. Fresh details, shared by a prominent hardware leaker, paint a picture of next-generation graphics cards potentially arriving with substantial increases in video memory and power requirements. If these rumors hold true, upcoming models like the RTX 5070 Super, RTX 5070 Ti Super, and RTX 5080 Super could offer notable upgrades over their standard counterparts, particularly appealing to users demanding more VRAM for high-resolution gaming, complex creative tasks, or advancing AI workloads.
Breaking Down the Blackwell Super Rumors
The source of this latest wave of speculation is the generally reliable Nvidia hardware leaker, @kopite7kimi, posting on X/Twitter. According to their shared preliminary specifications, Nvidia is gearing up to introduce at least three new desktop GPU models under the “Super” banner. These models are expected to sit alongside or potentially replace elements of the initially announced RTX 50 series lineup, all based on the new Blackwell architecture.
The notion of a mid-generation refresh isn’t new territory for Nvidia, with the “Super” designation previously signaling enhanced versions of existing cards, often featuring more cores, higher clocks, or increased memory. This leak suggests the RTX 50 Super lineup will follow a similar pattern, focusing primarily on boosting VRAM capacities and associated power envelopes.
RTX 50 Super Lineup: What the Leaks Say
The core takeaway from the rumored specifications points towards a notable uplift across the board compared to the non-Super variants. While changes to core counts appear relatively modest for some models, the jump in VRAM capacity stands out as a key differentiator. This could be a direct response to increasing demands from modern games and non-gaming applications alike, which are becoming increasingly hungry for large pools of high-speed memory.
The reported power consumption figures also see a significant rise. This aligns with the typical trend where increased performance, more cores, and faster, denser memory often necessitate higher Total Graphics Power (TGP). Consumers considering these rumored cards will need to factor in not just the potential purchase price, but also the costs associated with more robust power supplies and cooling solutions capable of handling the higher wattages.
Diving into the RTX 50 Super Specs
Let’s break down the preliminary leaked specifications for each rumored RTX 50 Super model, comparing them to their non-Super siblings based on the available information.
RTX 5070 Super: VRAM Boost and Power Bump
The entry point for the rumored Super refresh is said to be the RTX 5070 Super. Leaks indicate this card will utilize the GB205 GPU die, the same architecture underpinning the standard RTX 5070. It’s paired with a PG147-SKU65 PCB.
Regarding processing power, the RTX 5070 Super is rumored to feature 6,400 CUDA cores. This represents a modest increase of just over 4% compared to the 6,144 CUDA cores found in the standard RTX 5070. The real highlight here is the memory configuration. The leak suggests the 5070 Super will ship with a significant 18GB of GDDR7 memory. This is a substantial jump from the 12GB VRAM capacity of the standard RTX 5070, which was revealed to use 12GB of GDDR7 on a 192-bit bus with 28 Gbps bandwidth. The rumored 18GB version maintains the 192-bit bus interface but increases the total capacity, likely utilizing denser memory modules. The TGP for the 5070 Super is expected to rise to 275W, an increase of 25W over the 250W TGP of the non-Super model.
RTX 5070 Ti Super: High VRAM Potential
Moving up the stack, the RTX 5070 Ti Super is rumored to step up to the larger GB203 GPU, shared with the RTX 5080 series. It’s reportedly built on the PG147-SKU55 PCB.
This card is listed with 8,960 CUDA cores. While the non-Super RTX 5070 Ti specs weren’t detailed in the original leak, assuming it aligns with typical generational core count scaling or uses a cut-down GB203, the ‘Ti Super’ variant brings a significant VRAM increase into this performance tier. The leak points to a substantial 24GB of GDDR7 memory for the 5070 Ti Super. This memory is expected to run at 28 Gbps on a wider 256-bit bus interface. The TGP for this rumored card is listed at 350W, indicating a considerable power draw likely necessitated by the more powerful GPU and increased memory.
RTX 5080 Super: Max VRAM and Power
At the top of the rumored Super refresh lineup sits the RTX 5080 Super. This model is also expected to feature the GB203 GPU die, similar to the non-Super RTX 5080 and the rumored 5070 Ti Super.
According to the leaked details, the RTX 5080 Super is expected to maintain the same CUDA core count as the standard RTX 5080, listed at 10,752. The primary upgrade here, similar to the other Super variants, is the VRAM. The RTX 5080 Super is rumored to pack 24GB of GDDR7 memory on a 256-bit bus interface. This is a significant jump from the 16GB VRAM anticipated on the standard RTX 5080. The memory speed is also expected to be boosted to 32 Gbps, resulting in a memory bandwidth of 1 TB/s, a notable increase over the 960 GB/s rumored for the stock 5080. The power draw sees the largest increase in the lineup, with the RTX 5080 Super’s TGP potentially hitting 415W, a substantial jump from the rumored 360W TGP of the non-Super 5080.
VRAM: Why More Matters Now
The consistent theme across these rumored Super variants is the significant increase in VRAM capacity. This isn’t just a specification on a sheet; it has tangible impacts on real-world performance and future-proofing.
Modern gaming, especially at higher resolutions like 4K with demanding graphical settings and ray tracing enabled, can consume vast amounts of VRAM. Titles are increasingly pushing past the 12GB threshold previously common in high-end cards. Having 18GB or 24GB of VRAM provides a larger buffer, preventing performance bottlenecks and stuttering that can occur when a game tries to load more textures and data than fits in the available memory.
Beyond gaming, workloads such as professional video editing, 3D rendering, large-scale texture work, and particularly AI and machine learning tasks are extremely VRAM-intensive. Running large language models (LLMs) or complex AI inference models often requires large memory capacities to load the models efficiently. While Nvidia’s professional-grade GPUs (like the RTX Pro 6000 with 96GB or the RTX Pro 4000 with 24GB) are specifically designed and priced for these demanding tasks, offering 24GB on a consumer-focused RTX 5080 Super or 5070 Ti Super makes these cards much more appealing and capable for prosumers or researchers on a tighter budget.
The context provided by the announced 12GB RTX 5070 highlights the strategic decision behind these Super variants. While 12GB might suffice for many current gaming scenarios, Nvidia appears to recognize the need for higher VRAM tiers in the consumer lineup to address the most demanding uses and keep pace with evolving application requirements, positioning the Super cards as a significant upgrade path.
The Role of 3GB GDDR7 Modules
Achieving these higher VRAM capacities on memory interfaces like 192-bit or 256-bit often relies on using denser memory modules. Initial reports about the RTX 50 series suggested the use of 2GB GDDR7 modules. The rumored VRAM increases (18GB on a 192-bit bus, 24GB on a 256-bit bus) strongly imply that Nvidia plans to utilize 3GB GDDR7 memory modules for these Super variants.
For example, a 256-bit bus typically connects to eight memory modules (32 bits per module 8 modules = 256 bits). By populating these eight slots with 3GB modules instead of 2GB modules, the total capacity increases from 16GB (8 2GB) to 24GB (8 3GB). Similarly, a 192-bit bus connects to six memory modules (32 bits per module 6 modules = 192 bits). Using 3GB modules here results in 18GB (6 3GB) instead of 12GB (6 2GB). This technical detail underpins the rumored VRAM expansion across the lineup.
Performance Implications and Market Context
The rumored specifications, if accurate, suggest these Super variants will offer improved performance compared to their non-Super counterparts, driven by the increased VRAM and higher TGPs. More VRAM can directly translate to smoother frame rates and less stuttering in memory-bound scenarios. The higher TGP allows the GPU to potentially maintain higher clock speeds for longer durations under load, contributing to better performance.
Comparing Super Specs to Current & Announced Models
Putting these rumored Super specs into context with existing and announced cards is crucial. The non-Super RTX 5070 is announced with 12GB VRAM and 6144 CUDA cores. The rumored 5070 Super jumps to 18GB and 6400 cores. The rumored 5080 Super with 24GB VRAM and 10752 cores, sharing the GB203 die with the potential non-Super 5080 (rumored 16GB, 10752 cores), shows how VRAM is a primary upgrade vector for this refresh.
Compared to the previous generation, the RTX 4090 (24GB VRAM, 16384 shaders), even a rumored RTX 5080 Super doesn’t match the raw core count of the previous flagship. As discussed with the announced RTX 5070 potentially rivaling the 4090 via AI features like DLSS 4, the performance story for the 50 series and its Super variants will likely heavily involve AI acceleration rather than purely raw rasterization power versus older, higher-tier cards.
The Professional Card Divide
The inclusion of 24GB VRAM on consumer cards like the rumored RTX 5070 Ti Super and 5080 Super is notable in light of Nvidia’s professional GPU strategy. Professional cards often command high prices largely due to their extensive VRAM capacity. The RTX Pro 4000, featuring 24GB, was priced around $1,500. The recent emergence of a benchmark showing the $10,000 RTX Pro 6000 (based on GB202) outperforming even the RTX 5090 in gaming, largely due to its massive 96GB VRAM, underscores the value placed on memory in certain high-end applications.
Offering 24GB on a consumer card, even a high-end one like the 5080 Super, represents a strategic move to address the growing demand for substantial VRAM outside the traditional professional market, particularly from AI developers and prosumers. However, it also means Nvidia needs to carefully segment its market. Providing 24GB on the 5080 Super might set a ceiling for consumer VRAM below the top-tier 5090 (rumored to have 32GB) to avoid cannibalizing sales of even higher-VRAM professional cards, which can feature 48GB or more. The pricing of these Super cards will be critical; estimates for the 5080 Super have ranged from $1,000 to potentially $1,500, higher than many hoped but reflecting the increased VRAM and power.
Rumors vs. Reality: What to Expect
It is absolutely critical to remember that these specifications are based on leaks and rumors. While @kopite7kimi has a track record for accuracy regarding Nvidia hardware, until Nvidia makes an official announcement, all leaked details should be treated with a degree of skepticism. Specifications can change during the development process.
That said, the consistency of the VRAM focus across the rumored lineup suggests Nvidia is indeed looking to beef up the memory capacities in response to market trends. If these leaks are confirmed, the RTX 50 Super refresh could be a compelling option for users prioritizing VRAM-intensive workloads or seeking greater future-proofing. However, given Nvidia’s history, potential challenges with availability and pricing, especially at launch, remain a possibility.
Frequently Asked Questions
What are the key rumored specs for the Nvidia RTX 50 Super GPUs?
According to recent leaks, Nvidia’s rumored RTX 50 Super refresh includes models like the RTX 5070 Super, 5070 Ti Super, and 5080 Super. Key rumored upgrades include increased VRAM, potentially using 3GB GDDR7 modules. The RTX 5070 Super might feature 18GB VRAM and a 275W TGP. The RTX 5070 Ti Super and RTX 5080 Super are both rumored to sport 24GB VRAM on a 256-bit bus, with TGPs of 350W and 415W respectively. CUDA core counts also see slight increases on some models.
How significant is the VRAM increase rumored for the RTX 50 Super series for gaming and other tasks?
The rumored increase to 18GB and 24GB VRAM on the RTX 50 Super cards is highly significant. For gaming, more VRAM is crucial at high resolutions and with demanding settings like ray tracing, preventing performance issues when game assets exceed smaller memory buffers (like the 12GB on the standard RTX 5070). For professional workloads like video editing, 3D rendering, and especially AI/ML, larger VRAM capacity is essential for handling complex datasets and models, offering substantial performance benefits compared to cards with less memory.
Where did these RTX 50 Super leaks originate, and how reliable are they?
The primary source for these specific RTX 50 Super specifications is the hardware leaker @kopite7kimi, who shared the details on X/Twitter. @kopite7kimi has built a reputation for being a reliable source for Nvidia hardware leaks in the past. However, it’s important to treat any leaked information as preliminary and unconfirmed until Nvidia officially announces the products. Specifications can change between the leak and the final product release.
The rumored RTX 50 Super lineup, if realized with these specifications, signals a clear intent from Nvidia to address the growing demand for higher VRAM capacities in the consumer space. While the increased power draw is a notable factor, the potential for 18GB and 24GB of GDDR7 memory on mainstream to high-end consumer cards could make them very attractive for both demanding gamers and prosumers utilizing AI and creative applications. As always, the final performance, pricing, and availability will determine their true impact on the market, but these leaks provide a compelling glimpse into Nvidia’s potential next moves with Blackwell.
Word Count Check: 1176
