The most reliable Nvidia leaker on social media just reversed his own timeline. On September 21, 2026, Kopite7kimi, the source who set the earlier expectation of a second-half-2027 launch for Nvidia’s next GeForce generation, posted that the “GR20x” gaming GPU family has slipped to 2028. VideoCardz captured the post verbatim: “Unfortunately, GR20x is a product from 2028. That means they have postponed again.” He added that he plans to “minimize updates until there is definite news.”
That single post matters because it closes a gap that had opened just a week earlier. Nvidia has not made any official statement about a GeForce RTX 6000 series, its architecture, or a launch date. Everything below comes from leaks, supply-chain chatter, and reporting built on both, but three separate, independently sourced leakers now land on the same year for Nvidia’s next gaming lineup: 2028.
Kopite7kimi’s Reversal: From “2027H2” to 2028
Kopite7kimi has a track record worth taking seriously. Past leaks attributed to the account correctly previewed Nvidia product details ahead of official confirmation, which is why outlets treat the handle as one of the more dependable names in GPU rumors. On January 7, 2026, the same account posted that Nvidia’s next gaming architecture, codenamed GR20x, was targeting the second half of 2027, distinct from GR212, a related Rubin-family chip built for data center use rather than GeForce cards. PCWorld covered that claim at the time, citing an earlier report from Wccftech, and floated a possible announcement window around CES 2027 or late 2026.
Nine months later, that forecast is gone. Per VideoCardz’s report on the September 21 post, Kopite7kimi did not attach new technical detail to the reversal, only the blunt statement that the GPU family now belongs to 2028. Nvidia has not commented, and the account said it would hold off on further updates until something firmer surfaces. For a leaker who built credibility on precision, walking back a date by roughly a year, without new specifics to replace it, is itself a signal that something changed inside Nvidia’s planning rather than in the leaker’s guesswork.
A Rumor That Splintered, Then Reconverged, in Two Weeks
The most interesting part of this story isn’t the final answer, it’s how contested the timeline was just days before Kopite7kimi’s update. The dispute traces back to September 14, 2026, when the YouTube channel Moore’s Law Is Dead, citing a source described as inside Nvidia, claimed that “Gaming Rubin” had received an internal green light with a first-half-2027 target, meaning a launch window between January and June of that year. According to that account, the GeForce RTX 60 series had originally been planned for 2026, before Nvidia’s data-center Rubin products for AI customers took priority on manufacturing capacity.
That claim did not survive a full day of scrutiny. On September 15, 2026, leaker MEGAsizeGPU pushed back hard, arguing there was no supply-chain evidence supporting an early-2027 launch. As a point of comparison, Unboxholics reported that MEGAsizeGPU pointed to the RTX 5090, which had factory-level leaks surfacing as early as August 2024, roughly five months ahead of its early-2025 retail launch. No comparable manufacturing chatter exists yet for an RTX 6000-series card, which MEGAsizeGPU treated as strong evidence against a 2027 release of any kind.
Kepler_L2 and the AT2 Exception
A third voice had already weighed in before either of those posts. Leaker Kepler_L2, discussing the topic in an AnandTech forum thread, said AMD’s next high-end architecture, referred to as AT2, would be “the only new GPU to launch in 2027,” and that “everything else” from both AMD and Nvidia is being pushed back. For Nvidia specifically, that framing put the next GeForce generation outside the 2027 calendar entirely, well before Kopite7kimi’s own account came around to the same conclusion. Kepler_L2 did not publicly detail the source of that information or the underlying cause, beyond tying it to broader supply pressure that affects both companies at once. Shattered.io covered that original report in RDNA 5, Rubin GPUs Slip to 2028; AT2 Ships 2027, which lays out AMD’s side of the same delay in more detail.
What changed between then and now is corroboration. A single leaker calling a delay is a rumor. Three separately sourced leakers landing on the same year, after one of them spent months defending a different date, is closer to a pattern. That’s the actual news value in Kopite7kimi’s September 21 post: not a new detail, but the collapse of the one holdout position that had kept 2027 on the table.
Leaker Timeline: Who Said What, and When
The table below lines up every publicly reported claim about Nvidia’s next gaming GPU generation over the past nine months, drawn from the outlets that first reported each one.
| Source | Claimed timeline | Date reported | Reported by |
|---|---|---|---|
| Kopite7kimi (original) | Second half of 2027 (“2027H2”) | January 7, 2026 | PCWorld / Wccftech |
| RedGamingTech | RTX 6090 possibly Q1 2027 | Early 2026 | CanTech |
| The Information | Delays could push launch into 2028 | Mid-2026 | CanTech, multiple outlets |
| Kepler_L2 | 2028; AT2 is AMD’s only 2027 launch | September 2026 | AnandTech forum, via Unboxholics |
| Moore’s Law Is Dead | First half of 2027 (Jan-June) | September 14, 2026 | Unboxholics |
| MEGAsizeGPU | Not before 2028 | September 15, 2026 | Unboxholics |
| Kopite7kimi (revised) | 2028, “postponed again” | September 21, 2026 | VideoCardz |
Read chronologically, the pattern is clear: the 2027 case got weaker every time a new source weighed in, and the 2028 case got a fresh, higher-profile signature this week. CanTech’s own tracker of these claims, last updated September 12, 2026, reached a similar conclusion before Kopite7kimi’s reversal even landed: the most realistic scenario has an announcement around CES 2027 or late 2026, with retail availability sliding into “mid to late 2027” at the earliest and wider availability potentially stretching into 2028.
What “GR20x” Actually Refers To
Some of the confusion in this rumor cycle comes from codename overlap. GR20x is the internal label leakers use for Nvidia’s next gaming-focused GPU family, the one expected to eventually carry a GeForce RTX 6000 branding. It is distinct from GR212, which belongs to the Rubin CPX line built for data center and AI acceleration rather than consumer graphics cards. Both fall under the broader Rubin architecture that succeeds Blackwell, the platform underpinning the current RTX 50 series, but they are separate products on separate schedules, and only GR20x is relevant to gamers waiting on a next-generation graphics card.
That distinction matters for reading these leaks correctly. Nvidia’s data-center Rubin chips, aimed at AI customers paying far higher margins per unit, are reportedly getting manufacturing priority over GR20x. That prioritization is the mechanism every leaker points to, in one form or another, when explaining why a gaming GPU generation keeps sliding on the calendar. DropReference’s rumor roundup makes the same point: none of the leaked information so far includes core counts, memory bus width, or die sizes for GR20x, which is unusual this close to what was once billed as a 2027 launch and is itself a sign the design isn’t as far along in validation as Nvidia’s typical pre-launch leak pattern would suggest.
By contrast, the current-generation cards Nvidia does sell today, including the flagship listed on Nvidia’s own GeForce product page, have full, official specification sheets. The gap between “officially documented” and “leaker-only” is exactly what separates the RTX 50 series from anything branded RTX 6000 today, and it’s a gap that historically only closes a few months before a real launch, not a year or more ahead of it.
The Memory Shortage Sitting Underneath All of This
None of these leaks exist in a vacuum. A global shortage of GDDR7 and HBM memory, driven by AI data center demand, has been cited repeatedly as a factor squeezing next-generation gaming GPU schedules for both Nvidia and AMD. Shattered.io has tracked this shortage as it worsened through 2026, including Memory Chip Shortage: Stockpiles Fall Below 10 Days and comments from Intel’s own leadership warning that no meaningful relief is expected before 2028. That 2028 figure lining up with the leaked GR20x timeline is not a coincidence anyone in this rumor chain is treating lightly.
Memory isn’t the only constraint. TSMC’s advanced-node capacity is shared across every major chip customer building on similar process technology, and CanTech’s coverage of the rumor notes that current leaks point toward a 3nm-class node for Rubin-based gaming silicon, though Nvidia has confirmed none of the manufacturing specifics publicly. When AI accelerators and gaming GPUs compete for the same fabrication slots and the same memory suppliers, the accelerators win the priority fight, because the margins are larger and the customers are already committed to multi-year contracts.
RTX 6000 vs. the Current RTX 50 Series
It’s worth grounding this in what’s actually shipping today versus what’s still rumor. The table below separates confirmed facts about the current generation from the leaked, unconfirmed status of the next one.
| Attribute | RTX 50 series (current, confirmed) | RTX 6000 series (rumored, unconfirmed) |
|---|---|---|
| Architecture | Blackwell | Rubin (GR20x family) |
| Status | Shipping since early 2025 | No official Nvidia announcement |
| Retail pricing | Street prices for the flagship have surged well above original MSRP amid AI-driven demand, as Shattered.io reported | No pricing leaked or estimated by any cited source |
| Manufacturing node | TSMC 4NP-class process | Rumored 3nm-class, per CanTech, unconfirmed by Nvidia |
| Earliest leaked launch estimate | N/A, already shipping | H1 2027 per Moore’s Law Is Dead (disputed) |
| Latest leaked launch estimate | N/A | 2028, per Kopite7kimi, Kepler_L2, and MEGAsizeGPU |
Shattered.io covered the pricing side of that current-generation squeeze in more depth in RTX 60 Slips to 2028 as RTX 5090 Tops $5,000, which is worth reading alongside this piece for the market-pricing angle rather than the leaker-timeline angle covered here.
How This Lines Up With AMD’s Own Delay
Nvidia isn’t alone in this. Kepler_L2’s original AnandTech forum comments named AMD’s AT2 architecture as the one exception to the broader 2027 GPU drought, meaning AMD’s next high-end discrete GPU (widely expected to fall under an RDNA 5 or Radeon 10000-series banner for other products) is largely following the same delayed path as Nvidia’s GR20x. The practical effect is that gamers waiting for a meaningful generational leap from either company are looking at a market where almost nothing new arrives in 2027, with AT2 as the lone confirmed exception on AMD’s side.
| Company | Next major GPU family | Reported timeline | Key source |
|---|---|---|---|
| Nvidia | GR20x (rumored RTX 6000 series) | 2028 | Kopite7kimi, Kepler_L2, MEGAsizeGPU |
| AMD | AT2 | 2027 (on track, per Kepler_L2) | Kepler_L2, AnandTech forum |
| AMD | Everything else in the next lineup | 2028 | Kepler_L2 |
That leaves AT2 as one of the only genuinely new high-end GPU architectures with a 2027 date still attached to it from any credible leaker, a point Shattered.io detailed in its earlier coverage of the AMD side of this story.
Historical Context: Nvidia’s Generational Cadence Has Already Slowed
Nvidia’s consumer GPU releases have not followed a fixed clock for years, but the gaps between flagship architecture launches have been stretching. Blackwell followed Ada Lovelace after roughly two years, itself a longer gap than the transition before it. If GR20x lands in 2028 rather than 2027, the gap between Blackwell’s 2025 debut and Rubin’s gaming variant would stretch to roughly three years, the longest interval between major GeForce architecture changes in over a decade. That slower cadence tracks closely with the AI boom’s effect on fabrication and memory supply: the more profitable a chipmaker’s data-center business becomes, the less urgency there is to rush a lower-margin gaming lineup out the door on the old schedule.
Market Impact: What a 2028 Slip Means for Buyers
For anyone holding out on a GPU purchase in the hope that RTX 6000 cards are close, this rumor cycle is a signal to stop waiting. A 2028 timeline means the current RTX 50 series, and any Super or refreshed variants Nvidia releases in the interim, will remain the top of the consumer stack for at least two more years from today’s date. That has direct pricing consequences: with no imminent successor to create trade-in pressure, secondhand and retail prices for high-end RTX 50 cards have less reason to soften, particularly while AI-driven demand keeps competing for the same GPU dies and memory modules.
It also changes the calculus for PC builders and system integrators planning product roadmaps. A vendor betting on an early-2027 refresh cycle, based on the now-retracted Moore’s Law Is Dead claim, would have been planning around a launch window that no longer appears to exist according to the leaker who originally reported it. Retailers and OEMs generally plan inventory 12 to 18 months out, so a slip from H1 2027 to sometime in 2028 is the difference between one planning cycle and the next.
Competitive Landscape: Who Gains From a Slower Nvidia
A delayed Nvidia gaming lineup doesn’t automatically hand AMD an opening, since AMD’s own next-generation lineup outside AT2 appears to be sliding on a similar timeline, per Kepler_L2’s reporting. Intel is the more interesting case: its discrete GPU roadmap has moved more cautiously than either rival, and a longer Nvidia gap gives Intel’s Arc line more time to close the performance distance before facing a true next-generation Nvidia competitor. Meanwhile, Nvidia’s own AI-focused product lines, including its RTX Spark line for local AI workloads that Shattered.io covered in RTX Spark N1X Launches October, continue shipping on schedule, underscoring that this is a prioritization decision rather than a company-wide manufacturing failure.
The secondhand and refurbished GPU market is another likely beneficiary. With no near-term flagship refresh to devalue current-generation cards, resale prices for high-end RTX 50 hardware have less downward pressure than they would in a normal two-year cadence, which keeps used-market pricing elevated for longer than buyers might expect.
Predictions: Where This Rumor Goes From Here
- Expect at least one more leaker reversal or clarification before the end of 2026, given how quickly the timeline moved between September 14 and September 21 alone.
- Nvidia is unlikely to make any official statement about RTX 6000 naming or timing before CES 2027, even if that show ends up being an AI-hardware-first event rather than a gaming GPU reveal.
- Watch for a Super or Ti refresh of the current RTX 50 lineup in 2027 as a stopgap, a pattern Nvidia has used before when a full architecture change ran long.
- Memory pricing and availability, not raw GPU design work, will likely remain the single most-cited reason for further delays if the 2028 date slips again.
- AMD’s AT2 launching roughly on the 2027 schedule Kepler_L2 described would make it the most-watched high-end GPU release of that year by default, simply for lacking direct new competition from Nvidia.
What to Watch Next
The next real checkpoint is CES 2027, roughly 15 months from today’s date, where Nvidia has historically made major gaming-hardware announcements. Beyond that, keep an eye on TSMC capacity commentary and quarterly memory-supply reporting, since every leaker in this chain, from Kepler_L2 to MEGAsizeGPU to Kopite7kimi himself, points to the same upstream constraint rather than a design problem specific to GR20x. Until Nvidia says otherwise, the working assumption among the people who track this closest is now 2028, not 2027.
Related
- RTX 60 Slips to 2028 as RTX 5090 Tops $5,000 [2026]
- RDNA 5, Rubin GPUs Slip to 2028; AT2 Ships 2027 [2026]
- RTX 5090 Prices Double to $5,000 as AI Demand Bites [2026]
- Memory Chip Shortage: Stockpiles Fall Below 10 Days [2026]
- Intel’s Tan: Memory Up 7x, No Relief Until 2028 [2026]
Frequently Asked Questions
Has Nvidia officially confirmed the RTX 6000 series?
No. Nvidia has not announced a GeForce RTX 6000 series, its architecture name, specifications, or a release date. Every detail in circulation, including the GR20x codename, comes from leakers and third-party reporting, not from Nvidia itself.
Why did Kopite7kimi change his timeline?
The leaker did not give a detailed technical reason in the September 21, 2026 post reported by VideoCardz, only stating that GR20x had been “postponed again” and that he would limit further updates until there’s more definite information. The broader context, including comments from Kepler_L2 and MEGAsizeGPU, points to memory and manufacturing capacity constraints rather than a design-specific problem.
What is GR20x?
GR20x is the internal codename leakers use for Nvidia’s next gaming-focused GPU family, expected to eventually carry GeForce RTX 6000 branding. It’s separate from GR212, a Rubin-family chip built for Nvidia’s data center and AI products.
Is AMD’s next GPU generation also delayed?
Mostly. Per Kepler_L2’s reporting on AnandTech forums, AMD’s AT2 architecture remains on a 2027 timeline, but the rest of AMD’s next high-end lineup is reportedly moving to 2028 alongside Nvidia’s GR20x.
Should I buy an RTX 50 series card now instead of waiting?
If the current leaker consensus holds, a next-generation Nvidia GPU is roughly two years out from today. Buyers who need a GPU now have little practical reason to delay a purchase for an RTX 6000-series card at this point.
What’s driving the delay across both Nvidia and AMD?
Every source in this chain points to the same upstream issue: a global shortage of GDDR7 and HBM memory tied to surging AI data center demand, combined with limited advanced manufacturing capacity that AI accelerators are prioritized for ahead of consumer gaming GPUs.
When is the next likely update on this timeline?
CES 2027 is the next major event where Nvidia has historically made gaming-hardware announcements, though multiple leakers now expect that show to focus more heavily on AI hardware than on a new GeForce generation.
