Latest Rumors Suggest That NVIDIA Is Delaying Its RTX 50 SUPER Series Indefinitely, and the RTX 60 Series May Not Arrive Until Mid-2027

It’s possible that we might not see an RTX 50 series refresh anytime soon, thanks to the current memory chip supply shortage. More than a few were surprised at the absence of an RTX 50 SUPER series announcement at CES, but according to some industry insiders (via VideoCardz), things may be worse than expected. It seems that the AI monster, which NVIDIA helped create, has come hunting for the memory of its unreleased GPUs, and if true, the ongoing sentiment that the company, which once focused its resources on PC enthusiast hardware, is a step closer to abandoning them. A post, machine translated, on Board Channels is the basis for this latest speculation on what is happening with the GeForce RTX 50 SUPERs.

“There were earlier leaks saying that for the RTX 50 SUPER series, NVIDIA had already confirmed and informed AIC partners that the release would be delayed indefinitely. CES further supports this. There was no RTX 50 SUPER-related product showcase from NVIDIA. NVIDIA CEO Jensen Huang’s keynote also did not reveal or announce any new GPU products or related information.

The absence of a new-generation NVIDIA graphics card may disappoint some hardware enthusiasts, but this decision is not completely unexpected. The post suggests NVIDIA is mainly considering these factors:

1. Overseas demand for compute GPUs is surging, so more capacity is being reserved for compute GPU production, where profits are higher.
2. The rise in memory-related products may be a long-cycle issue. VRAM chips are not only expensive, but the cost also affects the global consumer GPU market share.
3. The most important point: there is currently no news of any new-generation AMD consumer GPU products launching in 2026. The current RTX 50 series can remain competitive for a long time, so releasing the SUPER refresh is less necessary.”

The current RTX 50 series is not immune to VRAM price increases

While many media outlets are content to report on the ongoing demand for current RTX 50 series graphics cards, most are overlooking a major point. That is, NVIDIA is still restricted to VRAM pricing as listed per its supplier contracts, and once those have expired, it will have to renegotiate them per current market conditions. That being said, it won’t matter if the RTX 50 SUPERs are released or not because the current product stack will then have its prices adjusted according to the new VRAM costs. This could mean we’ll see a $1,500 GeForce RTX 5070 or $1,000 RTX 5060.

As mentioned above, there’s no pressure on NVIDIA to produce new products right now as AMD hasn’t indicated it will be revealing anything new until at least 2027, and Intel has yet to formally announce its Big Battlemage B770 as well.

So many rumors, and none of them are good

The following is a breakdown, including details from one of our recent posts, from bad to worse.

1. NVIDIA may be considering the option to cease providing its partners with VRAM. AIB partners would then need to procure VRAM via their own channels, which could lead to some issues affecting prices pluys smaller companies might not be able to compete with larger companies whose orders wipe out supply.
2. Some current RTX 50 series could have reduced production. This was already suspected to happen with lower-tier cards due to rising VRAM costs, but as pointed out by the above quote from Board Channels, NVIDIA may also choose to focus its resources on compute-GPUs for their greater profitability. It was just last weekend that one overseas computer part seller reported that their supplier is already limiting which GPUs it can order and that it could not obtain any of the top-tier RTX 50 series (5070 Ti, 5080, or 5090).
3. RTX 50 SUPER memory upgrades could be too costly, right now. Back to the topic, as some rumors indicated the possibility of upping memory for the RTX 5070 Ti/5080, which use the same GPU die, to 24 GB GDDR7 and the 5070 to 18 GB, presumably using the newly available 3 GB modules. However, these premium modules are likely to come with a premium price, which would push these SUPERs beyond an acceptable pricepoint for consumers. Another roadblock for this launch, again mentioned already, is that NVIDIA could be eying the golden AI egg rather than less profitable enthusiast parts.
4. As reported this week, one other option NVIDIA could be looking into is reviving older GPU dies while pairing them with newer AI hardware and features. If NVIDIA were able engineer this feat while keeping in-house costs down, it might be able to create a stop-gap band-aid for consumers seeking some sort of upgrade path. However, there are plenty of unknowns with this approach, ranging from pricing and how much more powerful these older dies could be with newer AI magic sauce.

So, unless NVIDIA CEO Jensen Huang can pull a rabbit out of a hat, things are not looking for an RTX 50 SUPER launch anytime soon. However, even if it does happen, scalpers are likely already preparing to swipe said magic hat to resell for profits of their own.

Join the discussion in The FPS Review Forums...