Is Nvidia's Next-Gen RTX 60 Really Pushed Back to 2028? Inside the Rubin GPU Timeline Confusion

Gamers waiting on Nvidia's next flagship graphics cards have had a rough few months of rumor whiplash. According to one of the most closely watched hardware leakers, the GPUs once expected in the second half of 2027 may now not arrive until 2028 — and it's worth being upfront that none of this is confirmed by Nvidia. Here's what's actually circulating, who's saying it, and why it matters even as a rumor.

The Original Claim: Rubin Comes to Gaming

The story starts with kopite7kimi, a leaker on X (formerly Twitter) with a long track record of early, often-accurate Nvidia silicon details. Back in January, kopite7kimi reportedly said that Nvidia's next GeForce generation — informally dubbed "RTX 60" — would be built on "GR20x" dies derived from Rubin, the architecture Nvidia has so far positioned as a data-center and AI accelerator platform, not a gaming one.

That was notable because Rubin is the successor to Blackwell in Nvidia's AI roadmap, and it powers the company's Vera Rubin NVL72 systems aimed at hyperscalers and AI labs. Nvidia has, at various points, kept its data-center and consumer architectures separate — Volta, for instance, never made it into a GeForce card. So a claim that Rubin would double as the foundation for consumer RTX cards was treated as a genuine leak, not a foregone conclusion, and it still hasn't been corroborated by Nvidia in any official capacity.

According to reporting on the leak, the GR20x family would reportedly include GR202, GR203, GR205, GR206, and GR207 dies, succeeding the Blackwell-based GB202 through GB207 chips that power the current RTX 50 series. Some outlets peg GR202 as the likely flagship die, though it's worth noting the die naming itself is contested — other leaks have GR212 tied specifically to the Rubin CPX AI accelerator rather than gaming, and at least one outlet has flagged the entire naming scheme as unverified. Reporting here is doing a lot of hedging because the leakers themselves disagree.

Then the Timeline Slipped

Initially, kopite7kimi's reported timeline had RTX 60 arriving in the second half of 2027. More recently, that same leaker is said to have pushed the date out further, to 2028, in a reply disputing a rival leaker's claim. YouTuber and leaker Moore's Law Is Dead had claimed a first-half-2027 launch; kopite7kimi reportedly countered that the GR20x gaming lineup won't be ready until 2028.

That's a meaningful shift if accurate — it would mean well over three years between the RTX 50 series launch and its successor, longer than Nvidia's recent cadence. Moore's Law Is Dead's own sourcing reportedly claims the card once internally referred to as the "RTX 6090" was originally targeted for 2026, before being pushed back so Nvidia could dedicate more TSMC manufacturing capacity to AI chips instead. That claim comes from a separate source than kopite7kimi's and hasn't been independently verified, so treat it as a second, distinct rumor rather than confirmation of the first.

Why a Delay Is Plausible, Even Unconfirmed

Beyond the leaker chatter, there's a structural reason analysts find a delay believable: the ongoing memory shortage. Demand for AI accelerators has strained global supply of high-bandwidth and GDDR memory, and several reports this year have pointed to memory constraints as a real bottleneck across the GPU and PC industry, not just for Nvidia. If that shortage doesn't ease, it gives Nvidia a concrete business reason to prioritize AI data-center silicon — where margins are considerably higher — over a timely consumer refresh.

It's also consistent with broader industry rumblings. Reports around this year's Computex suggested that AMD's next architecture (UDNA, also called RDNA 5) and Nvidia's consumer Rubin cards could both slip toward late 2027 or even 2028, with board partners cited as the source. Jensen Huang himself has reportedly suggested RTX 60 might not land until 2027 at the earliest, while publicly emphasizing neural rendering as the more important near-term direction for gaming graphics — which, intentionally or not, lowers expectations for a near-term hardware leap.

What's Actually Confirmed vs. What Isn't

To be clear about where the solid ground is: Rubin itself, as a data-center architecture, is real and shipping. Nvidia confirmed Vera Rubin entered full production this year, and it's the backbone of upcoming AI supercomputing deployments. What is not confirmed is any of the following: that Rubin will be adapted for a consumer GeForce product at all, what it would be called, which die corresponds to which tier, or when such a card would ship. Every piece of that remains sourced to leakers reading supply chain signals and talking to board partners, not to anything Nvidia has stated publicly.

The Takeaway

If you're holding onto an RTX 40 or RTX 50 series card hoping for a near-term upgrade, the rumor mill's current best guess is that you may be waiting considerably longer than usual — but "current best guess" is the operative phrase. The timeline has already moved once in leaker reporting, the die names are disputed between sources, and Nvidia has given no official roadmap for a GeForce RTX 60 series. Until Nvidia says something itself, the only safe conclusion is that a Rubin-based gaming GPU is probably coming eventually, and nobody outside Nvidia reliably knows when.