RTX 5080 Super Release Date, Specs and Price: What Leaks Say

No, Nvidia has not confirmed the RTX 50 Super lineup — but a fresh wave of leaks says the RTX 5080 Super, RTX 5070 Ti Super, and RTX 5070 Super are still coming, with the latest reporting pointing to an October 2026 window. Leaked figures suggest a 24GB GDDR7 upgrade for the RTX 5080 Super and RTX 5070 Ti Super, and 18GB for the RTX 5070 Super, alongside modest CUDA core and clock bumps. Every number below comes from unofficial leaks and should be read as a rumor, not a spec sheet, especially given how many times this exact refresh has already slipped its rumored release window.

What's been leaked about the RTX 50 Super lineup

The RTX 50 Super rumor mill has been running for well over a year at this point. The broad strokes have stayed consistent across leakers and outlets: Nvidia is expected to refresh three cards in the middle of its Blackwell GeForce stack — the RTX 5080, RTX 5070 Ti, and RTX 5070 — with "Super" variants that carry more VRAM and slightly higher clocks, largely to answer criticism that the vanilla RTX 5070 and RTX 5070 Ti shipped with too little memory for 2026-era games. Tom's Guide reported that prolific leaker Moore's Law Is Dead was the source of the most detailed version of this leak, which also claims Nvidia plans to phase out the standard RTX 5080 and RTX 5070 Ti once the Super cards arrive.

That's consistent with how Nvidia handled the RTX 40 Super refresh in early 2024, when the RTX 4080 Super, 4070 Ti Super, and 4070 Super effectively replaced their non-Super predecessors at similar or lower prices. If Nvidia repeats that playbook, the vanilla RTX 5080 and RTX 5070 Ti could quietly disappear from Nvidia's own current RTX 50-series product lineup around the same time the Super cards launch.

RTX 5080 Super: leaked specs

The RTX 5080 Super is rumored to keep the same GB203 die as the standard RTX 5080 but with more of it enabled. Leaked figures cited by Tom's Guide and echoed by other outlets put the card at roughly:

• 10,752 CUDA cores in most versions of this leak — effectively unchanged from the standard RTX 5080's own core count, since the vanilla card already uses a near-complete GB203 die
• 24GB of GDDR7 memory, up from 16GB, reportedly running at 32+ Gbps
• A TDP of around 415W, roughly 60W higher than the standard RTX 5080's board power
• A rumored 9–16% performance uplift over the current RTX 5080, largely from the memory bandwidth increase rather than a big compute jump

If those numbers hold, the RTX 5080 Super would close a real gap for the card: 16GB has increasingly looked thin next to 24GB-and-up competitors at 4K with ray tracing and texture-heavy mods enabled. It's a similar story to how we compared the current card's memory ceiling in our RTX 5080 vs. RX 9070 XT breakdown, where VRAM headroom was one of the more contested points.

RTX 5070 Ti Super: leaked specs

The middle child of the rumored trio gets less coverage than its siblings, but the figures that have circulated point to:

• Around 8,960 CUDA cores, a modest step up from the standard RTX 5070 Ti
• 24GB of GDDR7 — the same memory capacity leakers expect on the RTX 5080 Super, effectively narrowing the gap between the two cards
• A TDP near 350W
• A rumored 7–11% performance uplift over the current RTX 5070 Ti

Because the RTX 5070 Ti Super and RTX 5080 Super are rumored to share the same 24GB memory pool, most of the differentiation between them would come down to CUDA core count and clocks rather than a memory-capacity gap — a departure from how Nvidia has usually spaced out its higher-end SKUs.

RTX 5070 Super: leaked specs

The RTX 5070 Super has the most consistently reported spec sheet of the three cards, likely because it comes from well-sourced leaker kopite7kimi as well as Moore's Law Is Dead. NotebookCheck's coverage lines up with other reporting on these figures:

• Roughly 6,400 CUDA cores, about 4% more than the standard RTX 5070's 6,144
• 18GB of GDDR7 memory at 28 Gbps on the same 192-bit bus, a 50% capacity jump from the current 12GB
• A TDP of about 275W, up from 250W
• Same GB205 die as the existing RTX 5070
• A rumored 8–12% performance uplift over the current RTX 5070

The RTX 5070 Super's story is almost entirely about memory. A 4% core-count bump won't move the needle much in benchmarks, but going from 12GB to 18GB is the kind of change that actually extends a card's useful life at 1440p and above — echoing the memory concerns we raised when comparing the current RTX 5070 to AMD's rival in our RX 9070 GRE vs. RTX 5070 vs. RTX 5060 Ti comparison.

Figures for the non-Super cards reflect Nvidia's official current specs. All Super-card figures are unconfirmed leaks and are presented as ranges where outlets disagree.

Release date: a saga of repeated delays

If there's one thing to be skeptical about here, it's the release date. This leak has already blown through several predicted launch windows. The earliest version of the story, sourced to Moore's Law Is Dead, pointed to a Q4 2025 launch, with Nvidia allegedly planning to discontinue the standard RTX 5080 and RTX 5070 Ti that October. That date came and went. Tom's Guide subsequently reported the cards had slipped into 2026 entirely, with a CES 2026 reveal looking likely at one point, before a separate report pegged the "most likely" window as March to May 2026, citing tipster Benchlife.

That window also came and went without a launch. By June 2026, leaker Zed__Wang reported the refresh was "back on track" for a launch before the end of the year, according to Club386's coverage — though notably without any specific month attached, and with a new wrinkle: an RTX 5060-class Super or "5060 12G" card reportedly joining the lineup. The most recent leaks tighten that to an October 2026 target, roughly a year after the refresh was first rumored.

Given that track record, treat "October 2026" as the current best guess rather than a locked date. Nvidia has offered no confirmation of a launch month, and this leak has been wrong about its own timing multiple times already.

Pricing: where leakers disagree

Pricing is the other area with real disagreement across outlets. Early reporting suggested Nvidia might hold Super pricing at parity with the outgoing non-Super cards — $549 for the RTX 5070 Super, matching the current RTX 5070. Other leaks, including figures attributed to RedGamingTech and cited by NotebookCheck, put the RTX 5070 Super closer to $599, a $50 premium over today's card. Recent leak roundups cluster around that higher figure alongside a rumored $1,149 price for the RTX 5080 Super, though some outlets have floated a wider $999–$1,199 range for that card. None of these figures should be treated as final — Nvidia typically doesn't lock in Super-series pricing until much closer to launch, and board-partner pricing tends to run above MSRP in the current GPU market regardless of what Nvidia lists.

Why Nvidia would launch a Super refresh now

Nvidia's Blackwell RTX 50 series launched at CES 2025, and a Super refresh roughly 18 months to two years later would track with the cadence Nvidia used for the RTX 40 series, which got its Super update about 14 months after the RTX 4090 debuted. The bigger pressure point, according to most of this coverage, is memory: the RTX 5070 and RTX 5070 Ti in particular have been criticized for shipping with 12GB and 16GB respectively at a time when competing cards and the demands of modern game engines have pushed past those numbers. A VRAM-focused refresh addresses that criticism without requiring a new GPU die, which is also why leakers expect Nvidia to reuse the existing GB203/GB205 silicon rather than introduce new chips.

It's also worth noting this wouldn't be an isolated move. Nvidia has continued iterating on its GeForce lineup between generations elsewhere too — see the compact RTX Spark's own release timeline for another example of Nvidia filling out the Blackwell family beyond its original CES announcements.

How the Super cards would stack up against AMD

AMD's current RDNA 4 lineup, including the RX 9070 XT and RX 9070 GRE, has made VRAM capacity part of its pitch against Nvidia's 12GB and 16GB cards. If the leaked 18GB and 24GB figures for the RTX 5070 Super and RTX 5080 Super hold up, Nvidia's refresh would largely erase that argument on paper, while likely keeping its usual advantages in ray tracing and upscaling via DLSS. Whether that's enough to shift the price-to-performance conversation depends heavily on where the Super cards actually land at retail versus AMD's response, if any.

Why this all still needs a grain of salt

To be direct about the sourcing here: none of this is confirmed by Nvidia. Every spec, price, and date in this article traces back to leakers — primarily Moore's Law Is Dead, kopite7kimi, and Benchlife — as reported by outlets including Tom's Guide, NotebookCheck, and Club386. Leakers with strong track records on GPU specs have sometimes gotten pricing and timing wrong, and this particular refresh has already missed at least two and arguably three previously rumored launch windows. Take the October 2026 date, the exact CUDA core counts, and the specific price points as the current consensus among leakers rather than settled fact.

What's next

The next real signal to watch for is a listing leak — retailer or customs filings, board-partner packaging, or a certification body entry that would tend to confirm a launch is imminent, similar to what preceded the original RTX 50-series launch. Until Nvidia says something directly, expect the release-date rumors to keep moving. We'll update this piece as Nvidia's own announcements — or the next round of leaks — arrive.

Written byMara Lindqvist

Mara Lindqvist edits the hardware desk. She covers graphics cards, processors, memory and storage, the foundries and chip designers behind them, and what the numbers on a spec sheet mean for people choosing a PC. Specifications in her stories come from manufacturer spec pages and datasheets.

More from Mara LindqvistAll Hardware coverage