Why the RTX 60 and RDNA 5 Delay to 2028 Changes When You Should Buy a GPU
Nvidia's next GeForce generation and most of AMD's RDNA 5 lineup are now expected to arrive in 2028, with only one AMD die holding a 2027 slot. The claim amounts to an RTX 60 and RDNA 5 delay of roughly a year against earlier expectations, and it comes from hardware leaker Kepler_L2 during a discussion on the AnandTech forums. A second leaker has since pointed the same way. Neither AMD nor Nvidia has published a consumer GPU roadmap for either year.
For anyone running a 2023 or 2024 graphics card and weighing an upgrade, the useful question is not who has the calendar right. It is whether waiting two years buys enough performance to justify sitting out a full cycle.
What the Leak Actually Claims
Kepler_L2's position is narrow. AMD's AT2 is the only new discrete gaming die on track for 2027, and it is built on the RDNA 5 architecture. The rest of AMD's stack, tracked under the codenames AT0, AT3 and AT4, moves to 2028. So does Nvidia's Rubin-based consumer family.
Nvidia's consumer Rubin parts are reported under five die designations: GR202, GR203, GR205, GR206 and GR207, all dated 2028. GR20x is the gaming derivative of the Rubin architecture and is expected to underpin the GeForce RTX 60 series. Kopite7kimi, a leaker with a long track record on Nvidia roadmaps, moved an earlier second-half 2027 schedule to 2028. Two sources now agree, though neither works for the companies involved.
| Product family | Codename / die | Reported timing |
|---|---|---|
| AMD RDNA 5 entry part | AT2 | 2027 |
| AMD RDNA 5 flagship and remaining dies | AT0, AT3, AT4 | 2028 |
| Nvidia RTX 60 (Rubin) | GR202, GR203, GR205, GR206, GR207 | 2028 |
If those dates hold, the gap between Nvidia's RTX 50 series and the RTX 60 series stretches to three years. AMD's Radeon RX 10000 family, built on RDNA 5 and tracked internally under the "GFX13" name, would skip a full cycle as well.
The asymmetry between the two companies is the part worth watching. AMD keeps a presence in the 2027 market with AT2 while Nvidia's consumer line goes quiet, which would put RDNA 5 branding in front of buyers a full year before Rubin cards appear. If AT2 lands as a mid-range part, that advantage stays in the volume segments rather than at the top of the stack.
Reported specifications for AT2 do not line up. One account describes a 40-compute-unit design, while another frames it as a roughly 64-CU successor to the Radeon RX 9070 XT. The discrepancy matters because it decides whether 2027 brings a mid-range card or something closer to the top of AMD's range.
Why the RTX 60 and RDNA 5 Delay Is a Capacity Story
The stated cause is manufacturing capacity. AI accelerator orders are absorbing foundry allocation and memory supply, particularly the HBM and GDDR7 that both data-center and consumer parts depend on. When a wafer slot or a memory stack can be sold into an AI accelerator at a higher margin, a gaming die competes from a weaker position.
That framing turns the slip into a business decision rather than an engineering failure. Nvidia's data-center business has grown far beyond its gaming segment, and AMD has aimed its own accelerator roadmap at the same buyers. Neither company needs a new consumer architecture in 2027 to defend its position, and both gain by keeping premium capacity pointed at AI.
The consequence for buyers is a longer plateau. Cards already on shelves stay current for longer, and the price-performance curve flattens. Existing owners get more useful life out of hardware they already own. Anyone waiting for a step change in efficiency per watt or per dollar has to wait roughly two more years.
What Could Move the Dates
Two variables sit under these timelines. The first is memory. GDDR7 and HBM supply constraints are named as a driver of the slip, which means an easing in memory availability could pull consumer launches forward without any change of intent at AMD or Nvidia. The second is the state of the AT2 plan itself. A single 2027 die with two conflicting spec accounts suggests a schedule still being shaped rather than locked.
Competitive pressure is the third factor, and the weakest of the three. Intel's discrete Arc line has not reached a scale that forces either company to accelerate. With no rival threatening share at the high end, there is little commercial reason to prioritize a gaming launch over an accelerator one.
Buy Now or Wait: Three Cases
Three groups of buyers face different answers.
- Owners of an RTX 40-series or RX 7000-series card have little reason to move. That hardware stays near the top of the consumer stack through 2027, and a 2028 replacement lands closer to the next real architectural jump.
- Anyone on an RTX 30-series, RX 6000-series or older card faces the opposite calculus. Two more years means two more years of declining frame rates in new titles, and a card bought today still holds resale value when 2028 hardware appears.
- High-end buyers should treat 2027 as a skip year. With no flagship RDNA 5 part and no RTX 60 card expected, there is no halo product to wait for.
The caveat is verification. This is a leak, not a launch schedule. Neither AMD nor Nvidia has confirmed product names, die counts or dates, and AMD's board partners have not been given a firm window. Roadmaps at this distance have shifted in both directions before.
A workable rule is to buy when the performance is needed rather than when a roadmap promises something better. The RTX 60 and RDNA 5 delay removes the strongest argument for waiting, which was that a new generation was only months away.
Why This Matters
Consumer graphics has run on a rough two-year cadence for most of the past decade, and that cadence is now set by AI demand rather than by gamers. The RTX 60 and RDNA 5 delay is the clearest evidence yet that gaming silicon queues behind accelerator silicon at the foundries. For PC builders, the practical result is a longer and calmer upgrade cycle, where timing a purchase around a launch window matters much less than it did a decade ago.
Related Articles
- AMD Gorgon Point Desktop Chips Leak with 60 TOPS AI Performance
- Graphics Card Prices Hit Record Highs as the AI Memory Crunch Reshapes Upgrades
- AMD Price Increase of About 10% Set for Q4 2026 as TSMC Wafer Costs Climb
✔Human Verified
Researched and cross-referenced against primary sources by the Bytevyte editorial team. This article was generated with the assistance of artificial intelligence and reviewed by the Bytevyte editorial team.