Nvidia’s most-followed hardware leaker just moved the goalposts on next-generation gaming GPUs again, and this time he’s not alone. On September 21, 2026, the leaker known as Kopite7kimi posted that the consumer version of Nvidia’s Rubin architecture, internally tracked as GR20x, will not ship until 2028. Tom’s Hardware reported the claim the same day, framing it as confirmation that gaming has taken a back seat while Nvidia funnels capacity toward data-center products. That framing lines up with the plain fact on the ground: as of August 2026, iTechGuides confirmed no RTX 60 model had been announced by Nvidia, with the RTX 50 series still standing as the current lineup, and Tech Insider separately noted that same month that no new Nvidia gaming GPU had shipped at all in 2026. The report matters less because it’s new information and more because of who is saying it: Kopite7kimi is the same source who, back in January, still expected RTX 60 cards in the second half of 2027. His own estimate has now slipped a full year.
This is not an isolated rumor anymore. It’s the third major leak in about a week pointing at the same 2028 window, following claims from Kepler_L2 and Moore’s Law Is Dead. For anyone tracking RTX 6090 release date chatter or wondering whether to buy an RTX 5090 now, that convergence is the real story: three independent sources, three different methods of reporting, one landing zone.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
What Kopite7kimi actually said on September 21
According to Tom’s Hardware’s translation of the post, Kopite7kimi wrote: “Unfortunately, GR20x is a product from 2028. That means they have postponed again.” The tech outlet Club386 reproduced the same wording in its own coverage of the delay. GR20x is the internal designation reportedly attached to the consumer, gaming-focused branch of Rubin, Nvidia’s next compute architecture after Blackwell. It is distinct from the data-center Rubin chips Nvidia has already previewed for AI customers, which remain on a separate and much faster-moving schedule.
Kopite7kimi has a long track record on Nvidia’s GPU roadmap, and that history is exactly why the September 21 post carries weight. Tom’s Hardware’s own report on the RTX 60 series delay to 2028 noted that this is a leaker who has previously called die configurations, memory bus widths, and launch windows well ahead of official confirmation. When a source with that record revises his own prior estimate a full year later, it reads less like speculation and more like an update based on new supply-chain information.
The timeline of the slide: from 2026 to 2028 in under a year
What makes the RTX 60 story unusual is how fast the target date has moved. Early industry chatter placed a Rubin-based gaming GPU launch in 2026. By late 2025, add-in board partners were reportedly skeptical of even that, with estimates already drifting toward late 2027 or 2028, according to Notebookcheck’s November 25, 2025 report citing AIB sourcing. Kopite7kimi himself told Tom’s Hardware in a January 7, 2026 report that he expected the RTX 60 family in the second half of 2027, a window that Tom’s Guide had independently floated that same month as a possible Q1 2027 launch. Neither estimate held: Tom’s Hardware reported in February 2026 that mass production of the RTX 60 generation might not even begin until 2028, and by July 2026 this site was reporting that Nvidia’s next generation looked set to skip 2027 entirely and land in 2028, with reporting suggesting Nvidia might lean on an RTX 50 Super refresh to cover 2027 instead of launching a new generation at all. And by mid-September 2026, two more leakers, Kepler_L2 and Moore’s Law Is Dead, had each independently pointed at 2028.
That’s four distinct downgrades in roughly ten months, each one pushing the expected launch further out rather than closer in. It’s the kind of pattern that tends to show up when a real supply constraint, not just conservative leaker hedging, is driving the schedule.
| Date | Source | Claim | Reported by |
|---|---|---|---|
| Nov 25, 2025 | AIB partner sourcing | Expectations already shifting toward late 2027 or 2028 | Notebookcheck |
| Jan 7, 2026 | Kopite7kimi (earlier estimate) | RTX 60 family expected in H2 2027 | Tom’s Hardware |
| Feb 5, 2026 | Supply chain reporting | Mass production seen sliding toward end of 2027, possibly 2028 | Tom’s Hardware |
| Jul 8, 2026 | Industry sourcing | RTX 50 Super refresh may cover 2027 instead of a new generation | Notebookcheck |
| Sep 15, 2026 | Moore’s Law Is Dead | RTX 60 targeted for H1 2027, delayed from an original 2026 plan | Notebookcheck |
| Sep 16, 2026 | Kepler_L2 | Rubin RTX 60 and most RDNA 5 GPUs pushed to 2028 | Notebookcheck |
| Sep 21, 2026 | Kopite7kimi (revised) | “GR20x is a product from 2028…postponed again” | Tom’s Hardware, Club386 |
Why three separate leakers now agree
Leaks converging on the same date isn’t proof by itself, but it does change the confidence level. Kopite7kimi, Kepler_L2, and Moore’s Law Is Dead draw from different corners of the supply chain, ranging from fab-adjacent contacts to AIB partner relationships, and they don’t always agree. When independent chains of sourcing land on the same year within a single week, it’s a stronger signal than any one leak on its own, which is why Tom’s Hardware treated the September 21 post as the lead story of the day rather than a minor update to an existing rumor.
It also lines up with a pattern this site has tracked since the first delay reports surfaced. A second leaker backed the same 2028 timeline just days before Kopite7kimi’s post, and that earlier convergence is exactly what made the September 21 claim land as confirmation rather than a fresh rumor.
The real driver: AI chips are eating Nvidia’s wafer allocation
Every report on the RTX 60 delay points back to the same underlying explanation: Nvidia’s data-center business is simply worth more per wafer than its gaming business right now, and advanced-node manufacturing capacity is finite. Tom’s Hardware characterized the situation directly, describing gaming as taking a back seat while the company prioritizes delivering data-center products. Moore’s Law Is Dead’s earlier reporting, cited by Notebookcheck on September 15, attributed an initial round of delays to Nvidia allowing AI products to consume as much advanced fabrication capacity as possible.
That tracks with Nvidia’s public financial priorities. Data-center GPUs and AI accelerators generate dramatically higher margins and volumes than consumer graphics cards, and every wafer allocated to a GeForce die is a wafer not building an AI chip that customers are reportedly waiting months to receive. None of the reporting reviewed here confirms a specific shortage tied to TSMC’s 2nm node; the claims instead point more broadly to Nvidia reserving advanced-node capacity for Rubin-generation AI products, a distinction worth keeping straight as the story spreads online.
GDDR7 and memory supply add a second bottleneck
Fab capacity isn’t the only constraint in play. Tom’s Hardware’s coverage also links the delay to memory availability, specifically constrained supply of denser 3GB GDDR7 chips that are relevant to both the RTX 50 Super refresh and any future RTX 60 cards. Memory supply problems have already shown up publicly: this site’s own reporting on the HBM4 shortage and falling DRAM inventory documented how thin memory buffers have become across the industry in 2026, and GDDR7 for consumer cards competes for some of the same upstream production lines as HBM4 for data-center accelerators.
Nvidia has acknowledged the memory side of this publicly, if not the launch-date side. In a statement Tom’s Hardware obtained regarding GeForce supply, the company said demand for GeForce RTX GPUs is strong and memory supply is constrained, adding that it continues to ship all GeForce SKUs and is working with suppliers to maximize memory availability. That’s a real acknowledgment of tight supply, even though it stops well short of confirming or denying any 2028 launch date.
What Nvidia has (and hasn’t) said
Nvidia’s official position on the GeForce RTX 60 timeline is silence. When Tom’s Hardware asked the company directly about the delay reports, Nvidia said it does not comment on rumors or unannounced products, which is its standard line for roadmap leaks. That non-answer leaves the 2028 timeline squarely in the leaker-and-analyst category rather than anything Nvidia has confirmed or denied on the record. It’s worth being precise about that distinction: everything in this story, including the GR20x designation, the 2028 window, and the RTX 6090 branding speculation, traces back to unofficial sourcing, not an Nvidia announcement or roadmap slide.
RTX 6090, GR20x, and a confusing naming picture
Names attached to this generation are still in flux. GR20x is the internal codename reportedly tied to the consumer Rubin die family, while RTX 6090 is the flagship product name that has circulated in roadmap discussions as the presumed successor to the RTX 5090, with RTX 6080, RTX 6070, and RTX 6060 filling out lower tiers in the same speculative framework. Specs chatter has filled some of that vacuum: Chinese outlet Baidu Baike reported in September 2026 that RTX 6090 rumors point to 21,760 CUDA cores, 32GB of VRAM, and a 512-bit memory bus, though none of those figures carry any official backing. The timeline itself remains contested too — ChainCatcher reported in October 2026 that RTX 6090 could still launch as early as H1 2027, a claim that sits awkwardly alongside Kopite7kimi’s 2028 post and underscores just how unsettled the leaker consensus really is. None of that naming, or those specs, or that timeline, is confirmed by Nvidia. As previously covered on this site, the RTX 60 series skipping a 2027 launch entirely was the first sign this generation would break from Nvidia’s usual cadence, and the branding uncertainty has persisted through every subsequent leak.
An architecture beyond Rubin, referred to in early roadmap discussion as Feynman, has also come up in reporting, but the available information places any Feynman-based product far enough out that a 2028 or 2029 window is speculative at best. For now, GR20x and RTX 6090 remain the only names with any leaker traction, and even those come with the caveat that Nvidia has changed consumer branding schemes before.
What this means for RTX 5090 owners and buyers
A later RTX 60 launch extends the shelf life of the current RTX 50 generation, and that has real pricing consequences. The RTX 5090 launched at a $1,999 MSRP in January 2025, but Notebookcheck reported it was effectively unavailable at that price at major retailers during launch week. Pricing swung wildly through 2025: distributor PowerGPU was quoting roughly $3,050 to $3,100 for entry-level cards in March 2025, according to Notebookcheck, before Nvidia cut European pricing that August, with the RTX 5090 dropping from around €2,320 to €2,099, a 9.9% reduction as availability improved.
That relief didn’t last. Tom’s Hardware reported on August 3, 2026 that GPU prices remained elevated, and with a next generation now pushed potentially into 2028, buyers face at least another year, and possibly two, without a new flagship architecture to wait for. This site’s own tracking of RTX 5090 pricing climbing alongside the RTX 60 delay and the broader GPU price relief being pushed to 2028 amid a DRAM price surge both point the same direction: fewer reasons for retailers to discount a card with no near-term successor, and continued upward pressure from memory costs on top of that.
AMD’s RDNA 5 hits the same wall
If Nvidia were delaying alone while AMD shipped on schedule, that would open a real competitive window. The leaks don’t support that scenario. Notebookcheck’s September 16 report citing Kepler_L2 stated that Rubin-based RTX 60 cards and most of AMD’s RDNA 5 lineup would both land in 2028, with only one possible AMD product arriving earlier. AMD has not confirmed a 2028 timeline for its next Radeon architecture any more than Nvidia has confirmed one for RTX 60, but the leaked picture suggests both companies are wrestling with the same advanced-node capacity crunch rather than one racing ahead of the other. That’s consistent with this site’s earlier coverage of AMD’s current-generation GPU sales gains, which found AMD picking up market share largely on pricing and availability of existing cards rather than any next-generation advantage.
For gamers hoping a delay from one vendor would hand the other a decisive lead, the leaked timeline suggests something closer to a shared pause. Whoever ships first, even by a few months, could gain an outsized marketing advantage simply because the wait has stretched so long.
Does Intel get a competitive window?
Intel is the wildcard in this story, and it’s a genuinely open question rather than a confirmed opportunity. Reporting on Intel’s future Celestial architecture exists, but none of the sources reviewed for this story establish a firm 2027 or 2028 retail launch date for it. If Intel manages to ship a meaningful Arc successor while Nvidia and AMD are both stuck on 2028 roadmaps, it could carve out attention in the mid-range segment where Intel has already been competitive on price. But that’s a possibility built on an absence of confirmed Nvidia and AMD launches, not on any specific Intel commitment reported alongside the RTX 60 delay news.
Historical context: Nvidia’s generation cadence, broken
Nvidia’s consumer GPU releases have followed a roughly two-year rhythm for most of the last decade. Ampere, the RTX 30 series, arrived in 2020. Ada Lovelace, the RTX 40 series, followed in 2022, a two-year gap. Blackwell, the RTX 50 series, launched in January 2025, a gap of roughly two and a half years. If Rubin-based RTX 60 cards land in 2028 as the latest leaks suggest, that would stretch the gap to around three years, the longest span between consumer GeForce generations in recent memory. Wikipedia’s entry on the RTX 50 series documents that January 2025 launch date, which anchors the two-and-a-half-year comparison point used throughout this story.
| Series | Architecture | Launch | Gap from prior gen |
|---|---|---|---|
| RTX 30 | Ampere | 2020 | — |
| RTX 40 | Ada Lovelace | 2022 | ~2 years |
| RTX 50 | Blackwell | January 2025 | ~2.5 years |
| RTX 60 (leaked) | Rubin / GR20x | 2028 (unconfirmed) | ~3 years |
Tom’s Hardware explicitly framed the possible 2028 launch as a break from Nvidia’s near two-year consumer cadence, and the table above shows why: even the most conservative reading of the leaks points to the longest gap of the modern RTX era.
Market impact beyond gaming GPUs
The ripple effects extend past desktop gaming cards. Nvidia’s professional and workstation lines, including cards like the RTX PRO 5500 with its 84GB memory configuration, sit on adjacent product roadmaps that could shift alongside any Rubin consumer delay. The mobile side is in motion too: VRLatech reported in June 2026 that Nvidia’s RTX Spark laptops, built around unified memory configurations of up to 128GB, were on track for a fall 2026 launch, giving Nvidia a high-memory mobile story to lean on while the desktop RTX 60 generation keeps slipping. Regional retail markets are already feeling pricing pressure tied to the broader GPU supply story: this site’s coverage of Akihabara GPU prices jumping amid RTX PRO 6000 shortages showed the same dynamic playing out at the professional tier, where enterprise AI buyers are competing with gamers and creators for limited supply.
Niche markets are affected too. Sim racing rigs, which typically demand top-tier GPUs for high refresh-rate triple-monitor setups, have already seen price pressure documented in this site’s report on sim racing GPU prices climbing toward $5,000 for RTX 5090-class hardware. A longer wait for RTX 60 removes any near-term relief valve for that segment of enthusiast buyers.
What industry voices are saying
The clearest first-hand account remains Kopite7kimi’s own post, translated by Tom’s Hardware as: “Unfortunately, GR20x is a product from 2028. That means they have postponed again.” Club386’s independent write-up of the same leak reproduced identical wording, which is notable since the two outlets appear to have translated or sourced the post separately and arrived at the same phrasing.
Notebookcheck’s reporting on the parallel Kepler_L2 leak summarized the claim plainly, noting that the source states Nvidia’s Rubin-based RTX 60 series, as well as AMD’s RDNA 5 GPUs, will launch in 2028. And Tom’s Hardware’s own headline on the story, “Nvidia’s next-gen RTX 60 GPUs might not be released until 2028,” captures the outlet’s framing of the reported consensus forming around that year. Taken together, these three independent write-ups describe the same shift from three separate vantage points, which is a meaningfully different situation than a single leak circulating unchecked.
Predictions: what happens between now and 2028
- Expect an RTX 50 Super refresh, rather than a genuine new architecture, to be Nvidia’s primary consumer GPU news through 2027, extending Blackwell’s commercial life instead of replacing it.
- RTX 5090 pricing is unlikely to soften meaningfully while it remains the flagship with no announced successor, especially if DRAM and GDDR7 costs stay elevated into 2027.
- Watch for additional leaker convergence rather than a single dramatic reveal; expect more sources lining up behind 2028 in the coming months as more supply-chain contacts weigh in, rather than a sudden Nvidia announcement pulling the date forward.
- AMD’s RDNA 5 timeline will likely track Nvidia’s delay closely, given both companies reportedly face the same advanced-node capacity constraints, so a major AMD leapfrog moment looks unlikely before 2028.
- Nvidia’s official commentary will probably stay limited to supply acknowledgments, similar to its GeForce memory statement, without a formal roadmap confirmation until much closer to an actual launch window.
Frequently asked questions
When will the Nvidia RTX 60 series be released?
No official release date exists. As of September 21, 2026, leaker Kopite7kimi placed the consumer Rubin-based lineup, internally tracked as GR20x, in 2028, a claim echoed by separate leakers Kepler_L2 and Moore’s Law Is Dead in reports published by Notebookcheck earlier the same week. Nvidia has not confirmed any date.
Who is Kopite7kimi?
Kopite7kimi is a well-known hardware leaker frequently cited by outlets including Tom’s Hardware for Nvidia GPU roadmap information. The leaker’s September 21, 2026 post revised an earlier estimate, reported by Tom’s Hardware in January 2026, that had placed RTX 60 in the second half of 2027.
What is GR20x?
GR20x is the internal codename reported for the consumer, gaming-focused implementation of Nvidia’s Rubin architecture, distinct from the data-center Rubin chips already previewed for AI customers.
Is the RTX 6090 name officially confirmed?
No. RTX 6090 has circulated in roadmap discussions as the likely flagship name for the next Nvidia consumer generation, alongside speculative RTX 6080, RTX 6070, and RTX 6060 tiers, but none of this naming has been confirmed by Nvidia.
Why is Nvidia reportedly delaying gaming GPUs?
Reports point to Nvidia prioritizing advanced manufacturing capacity for higher-margin data-center AI products, combined with constrained GDDR7 memory supply that also affects the RTX 50 Super refresh. Nvidia has publicly acknowledged strong GeForce demand and constrained memory supply, without addressing a specific launch date.
Will AMD’s next-generation GPUs launch before Nvidia’s?
Leaked reporting suggests most of AMD’s RDNA 5 lineup is also headed for 2028, alongside RTX 60, according to Kepler_L2’s claim as reported by Notebookcheck, with only one possible AMD product arriving earlier. Neither company has confirmed a date.
Should I buy an RTX 5090 now or wait for RTX 60?
With multiple leakers now pointing to a 2028 launch, buyers who need a GPU today face a long wait with no confirmed near-term alternative. Prices have fluctuated significantly since the RTX 5090’s January 2025 launch, and elevated pricing has persisted into 2026 according to Tom’s Hardware reporting, so timing a purchase around a specific discount window carries real uncertainty either way.
Has Nvidia responded to the delay reports?
Nvidia told Tom’s Hardware it does not comment on rumors or unannounced products. The company has separately acknowledged strong GeForce demand and constrained memory supply in an unrelated statement about current-generation availability.
Related Coverage
- Second Leaker Backs RTX 6000 Delay to 2028 [2026]
- RTX 60 Delay to 2028: AMD Radeon Caught in Same Freeze [2026]
- RTX 6090 Release Date Slips to 2028: 3-Year Wait [2026]
- AMD Hits $1T: Wall Street Splits $76 on Next Move [2026]
- AMD Joins $1T Club, Stock Jumps 10% Past Intel [2026]
