Is Nvidia's Next GPU Generation Slipping to 2028? Leakers Can't Agree on the RTX 60 Timeline

Nvidia hasn't confirmed a single detail about its next graphics card generation, yet the leaker community is already in open disagreement about when it's actually coming. The cards in question, widely expected to carry the "RTX 60" branding and sit on Nvidia's upcoming Rubin architecture, are reportedly the successor to today's RTX 50 series. According to competing reports circulating through September and into October, the launch window has swung between the second half of 2027 and, more pessimistically, 2028 — depending on which leaker you ask this week.

None of this is official. Nvidia has not announced an RTX 60 series, has not confirmed the Rubin name for consumer GPUs, and has not committed to any date. But the pattern of claim-and-counterclaim is itself a story worth tracking, because it shows just how unsettled even well-connected leakers are about Nvidia's roadmap right now.

Where the "2027" Claim Came From

The timeline leak traces back to Moore's Law Is Dead (MLID), a YouTube-based hardware leaker with a mixed but notable track record on Nvidia roadmaps. MLID reported that an Nvidia source told him the company is "aiming for H1 2027" for the RTX 60 launch. Later, reportedly on September 23, MLID said he went back to the same source for more detail and came away with a claim that the "Rubin Client" silicon — the consumer-facing die that would presumably power the RTX 6090 and its siblings — is being targeted for a 2027 release.

That's a notably more optimistic timeline than what some other leakers have floated, and it set up a public disagreement.

Kopite7kimi's Pushback

Kopite7kimi, a leaker generally regarded in the GPU enthusiast community as one of the more reliable sources on Nvidia's internal roadmaps, reportedly undercut MLID's claim just days earlier, suggesting the Rubin-based gaming lineup had slipped toward 2028 rather than 2027. Kopite7kimi's own earlier estimate — before that apparent walk-back — had pointed to the second half of 2027, which is itself not far from MLID's number but illustrates how fluid these projections are even from a single source over time.

It's worth sitting with that for a second: two of the most-cited leakers in this space are not citing the same internal roadmap language, and even the "more reliable" one has reportedly shifted his own estimate. That's a signal that Nvidia's actual internal planning may still be moving, rather than that any one leaker has nailed it down.

The Skeptics Are Loud Too

Not everyone is convinced there's a firm 2027-vs-2028 debate worth having at all. Leaker MEGAsizeGPU reportedly dismissed the whole timing discussion as "fake-as-hell," pointing instead to Nvidia's ongoing supply chain constraints — chiefly the crunch on advanced packaging and leading-edge wafer capacity that has defined the AI-chip boom — as the real bottleneck that will determine when consumer Rubin GPUs can actually ship in volume. Outlet VideoCardz, which maintains its own sourcing on Nvidia's roadmap, reportedly said it hadn't picked up fresh signals on the gaming Rubin lineup specifically, which is itself a data point: if a launch were genuinely locked for H1 2027, it's a little surprising that more outlets aren't converging on consistent intel by now.

What the Specs Rumors Say (Treat Cautiously)

The specification chatter around a prospective RTX 6090 is, if anything, even softer than the timing rumors. The architecture is expected to carry forward Nvidia's codename convention with a "GR20x" family for gaming Rubin parts — GR202 reportedly as the flagship, with GR203, GR205, GR206 and GR207 arrayed below it as high-end, upper-mid, mid and budget tiers respectively, according to kopite7kimi's lineup. One widely circulated figure claims a roughly 28,000 CUDA-core count, but that number appears to originate from die analysis of "Rubin CPX" — understood to be a separate, data-center-oriented part (reportedly GR212), not the consumer flagship. Conflating the two would be a mistake, and several outlets covering this have flagged exactly that confusion already taking hold online.

The more consistently repeated claim — though still unconfirmed — is that a flagship RTX 6090 would carry 32GB of GDDR7 memory on a 512-bit bus, which would track with Nvidia's historical pattern of widening the memory bus at the top of a new generation.

Why This Matters Beyond the Spec Sheet

For UAE and wider Gulf buyers and businesses, the practical stakes are less about bragging rights over CUDA core counts and more about planning. The region's AI infrastructure build-out — from G42's data center ambitions to the broader Gulf push into sovereign AI compute — leans heavily on Nvidia's data-center silicon, and the consumer Rubin cards share architectural DNA with those parts. A slip from 2027 to 2028 on the gaming side doesn't necessarily delay data-center Rubin parts on the same timeline, but the public disagreement among leakers underscores that Nvidia's broader Rubin rollout, across both consumer and enterprise lines, still hasn't settled into a predictable cadence that outside observers can confidently bank on.

The Bottom Line

Strip away the specific numbers and what's left is a familiar pattern in GPU leak culture: multiple leakers, multiple "sources," and no agreement. Nvidia itself has said nothing public about an RTX 60 series, a Rubin gaming architecture, or any 2027/2028 date. Until the company holds an actual roadmap presentation — something it has historically done well over a year ahead of silicon shipping — treat every date in this story, including the ones attributed to "reliable" leakers, as provisional at best. The only safe prediction right now is that there will be more conflicting leaks before there's an official Nvidia keynote slide to settle it.