Last Updated on
We’re here to go over all Nvidia RTX 50-series rumors so that you are clued up on every exciting development. Now that we’re firmly in March 2024, we are creeping ever closer to the fabled release window as Team Green typically releases a new graphics card generation in two-year spans, so there’s not long now to learn about the ace up its sleeve. There are a ton of credible rumors and leaks to go through, so let’s get into all the details.
Nvidia RTX 50-series: Release date rumors
The Nvidia RTX 50-series, codenamed “Blackwell,” is generating significant buzz, with industry experts and enthusiasts alike anticipating its arrival. According to a video from Moore’s Law is Dead, citing an anonymous Nvidia source, we could see a Q4 2024 launch. This falls in line with Nvidia’s traditional launch patterns for new GeForce RTX series graphics cards.
This potential release window is consistent through the Chinese hardware website UDN which explains that the RTX 5090 and RTX 5080 will be released at the end of 2024. Both MSI and Gigabyte are mentioned in the article as two partners that allegedly will have AIBs ready at the end of this year, which is consistent with when we saw the debut of the RTX 40 series two years ago which also released with the two flagship models first as the mid-range and mainstream came later.
However, it’s important to note that this timeline isn’t set in stone. Market dynamics, particularly the performance and reception of AMD’s upcoming RDNA 4 architecture, could significantly influence Nvidia’s roadmap. Expect Nvidia to closely monitor the competition in the crucial 2024 sales period. This means that, while late 2024 is the current target, flexibility is built into the release strategy.
Nvidia is playing a calculated game, ready to optimize the RTX 50-series launch to maximize its impact on a changing GPU architecture landscape. This cautious approach highlights Nvidia’s continued dedication to providing cutting-edge graphics solutions, including potential leaps in ray tracing performance, while considering broader market forces. To throw some contrast in though, one rumor does suggest a release date as early as September 2024, just before we hit Q4.
Nvidia RTX 50-series: Specs leaks
The Nvidia RTX 50-series, set to follow the Ada Lovelace architecture, will be named “Blackwell.” With rumors swirling, the most concrete details suggest this next-generation GPU will be crafted on TSMC’s advanced 3nm process node.
According to TSMC, the 3nm process node is expected to deliver up to a 15% performance boost over the 5nm process at the same power consumption, with die sizes shrinking by an estimated 42%. This could significantly impact the design of Blackwell’s silicon, making for a more efficient yet powerful GPU, especially compared to the large 609 mm² die of the GeForce RTX 4090. Who wouldn’t love smaller GPUs that are stronger?
Furthermore, renowned leaker Kopite7kimi’s insights further enhance our understanding of what the RTX 50-series might offer. According to these leaks, Blackwell GPUs could introduce support for DisplayPort 2.1 and HDMI 2.1, enhancements not present in the Lovelace lineup.
Additionally, the series is expected to utilize next-gen GDDR7 memory, with a maximum memory bus width of 384-bit, pointing towards substantial improvements in bandwidth and overall performance. This would mark the first time that Nvidia has upgraded the video memory in its graphics cards since the GTX 1000, the last line to use GDDR5X memory as standard before GDDR6 was ushered in with the RTX 2000 series in 2018. With faster VRAM, comes far higher bandwidth, and better performance.
| Feature | Nvidia RTX 50-series |
|---|---|
| Architecture | TSMC 3nm |
| Process node | Blackwell |
| Chip | GB202, GB203, GB205, GB206, GB207 |
| Memory type | GDDR7 |
| Maximum bus width | 384-bit |
Adding PCIe 5.0?
Another intriguing rumor is the potential inclusion of the PCIe 5.0 interface, though its necessity for mainstream graphics cards remains debatable. Nonetheless, the RTX 50-series will probably maintain the use of the 16-pin power connector, likely transitioning to the revised 12v-2×6 version for enhanced safety. This move is particularly noteworthy in light of the recent controversies surrounding the 12VHPWR connector, with Nvidia largely attributing issues to user error while the community eagerly awaits a more robust solution.
Adding to the performance expectations for the Nvidia RTX 50-series, insights from Moore’s Law Is Dead suggest a nuanced picture of potential gains. The source mentioned to the YouTuber indicates that while we may not see a groundbreaking leap in rasterization performance from Ada to Blackwell, the improvements will still be noteworthy.
Specifically, the expectation is set for a 30% to 50% performance increase for flagship models. This range, while significant, highlights a more incremental improvement compared to previous generational leaps, particularly from Ampere to Ada.
The conversation also touches on strategic considerations, hinting that Nvidia could theoretically push for more aggressive performance enhancements for the RTX 5090 if it felt pressured by competition. However, with AMD reportedly dialing back its ambitions in the high-end GPU market for the next generation, Nvidia might not face the kind of competitive push that necessitates a drastic performance uplift.
This context suggests that Nvidia’s position as a leading provider of high-end graphics cards may remain unchallenged for the next couple of years, potentially influencing its approach to performance improvements across the RTX 50 series.
While we’ve known that Blackwell could be using GDDR7 video memory for some time, what’s been breaking news since has been the alleged bandwidth with 28 Gbps effective memory (via WCCFTech). This source claims that we could see the top-end GB202 and GB203 die cards, possibly RTX 5090 and RTX 4080, with memory bandwidths of up to 512-bit, which is considerably larger than the 384-bit of the current Ada leader, the RTX 4090.
Interestingly, the leaker @kopite7kimi on Twitter, who broke the story, claims that there could be a bandwidth of up to 1.8TB/sec which would be a drastic improvement over the RTX 4090 which has a bandwidth of 1,008 GB/sec. Cast your mind back to 2022 and the RTX 3090 Ti was the first GPU to push into the 1,000 GB/sec territory, now, it’s been rumored the Blackwell flagship could be faster by 78% (via Videocardz).
Speaking of Blackwell, Nvidia officially announced this architecture back in March to “power a new era of computing”. The B200 dual die is reported to be 4x faster than Hopper architecture for AI-accelerated tasks. As to how this larger die will be integrated into graphics cards for gaming remains to be seen, however, considering the innovations brought in by DLSS 3’s Frame Generation, it looks like we’ll be moving away from native performance and more to how AI can up-res games for higher FPS.
Mid-range and entry-level cards rumors
For mid-range and entry-level cards within the 50-series, the performance boosts are expected to be more modest. These segments typically witness smaller generational improvements, aligning with the broader trend of focusing on flagship models for the most substantial upgrades. This stratification ensures that while enthusiasts and professional users can anticipate notable enhancements at the high end, the average consumer will still see improvements, albeit at a scaled-down magnitude.
Nvidia RTX 50-series: Price predictions
Speculating on the Nvidia RTX 50-series pricing is a tricky endeavor, given the current market dynamics and Nvidia’s recent pricing strategies, which many would describe as premium. With the RTX 4090’s price surging beyond $2,000 in the market, driven by high demand, especially for AI-focused GPUs, there’s concern that Nvidia might continue this trend with the upcoming Blackwell series. The flagship RTX 5090, for instance, could potentially hit close to a $2,000 price point at launch if Nvidia decides to further escalate its pricing.
However, this approach brings its own set of challenges, particularly in maintaining competitiveness with AMD. Despite the potential for higher prices, there’s an underlying expectation—or perhaps a hope—that Nvidia will acknowledge the ceiling for sustainable pricing strategies. The tech community is keenly watching for signs that Nvidia will adopt a more balanced approach to pricing its next-generation GPUs, ensuring that they remain attractive to a broad audience without compromising on the cutting-edge performance that enthusiasts expect.
While the top-tier models like the RTX 5090 might command premium prices, the ripple effect across the lineup could see overall price increases. Yet, the imperative to stay competitive, especially in segments where AMD remains a strong contender, might temper Nvidia’s pricing ambitions. It’s a delicate balance between capitalizing on advanced technology and maintaining market dominance, making the RTX 50-series pricing a subject of much speculation and anticipation as we edge closer to its unveiling.
Nvidia RTX 50-series: Final word
While firm details remain elusive, speculation and leaks surrounding the Nvidia RTX 50-series ‘Blackwell’ point to a promising new chapter in GPU technology for gamers and enthusiasts. The transition to the cutting-edge 3nm process promises efficiency gains and potentially raw performance improvements. However, it’s essential to set realistic expectations.
Remember, the already powerful Nvidia GeForce RTX 40-series set a high baseline. While incremental improvements on multiple fronts are likely, expecting a groundbreaking 30-50% generational leap in all benchmarks may not match the normal pace of hardware development. Nvidia will undoubtedly aim for advancements with its Blackwell architecture but consider them significant optimizations for a new series of GPUs more than a revolution.
Ultimately, Nvidia’s RTX 50-series release date rumors hinge not only on technological readiness but on the broader competitive landscape. Balancing market forces and delivering truly compelling products are essential for Nvidia’s continued success. Stay tuned for more concrete information as the potential launch window grows closer. Time will tell how the RTX 50-series GPUs will rival the best GPUs on the market today.
Will Nvidia make a 50-series?
While the Nvidia RTX 50-series has yet to be announced, it’s a strong possibility that Team Green will make another GPU generation in league with the previous three lines, especially as the GTX moniker was recently retired. The company typically brings out a new graphics card series every two years and the last one was in Q3 2022.
How much could the RTX 5090 cost?
Considering the RTX 4090 retails for $1,599 and the previous-generation BFGPU sold for $1,499, then it’s possible that the RTX 5090 could sell for $1,699 in keeping with the $100 increase generationally. However, this has yet to be confirmed by Nvidia.
