NVIDIA RTX 60 series: the full rumored lineup, tier by tier

“RTX 60 series” is already a real search — so here's the honest version of what's known. NVIDIA's next GPU generation is rumored to be called Rubin, arriving somewhere between 2027 and 2028, in a lineup that mirrors today's RTX 50 stack. Below is every tier that has leaked, what's claimed for it, and — the part most articles skip — how confident anyone can actually be about each one. If you want a card you can buy today, jump to our best graphics cards of 2026 instead.

The rumored RTX 60-series lineup

Here's the full stack as it stands in the leaks. Read the confidence column as carefully as the specs — it's the difference between “multiple leakers agree” and “someone guessed a codename.”

No CUDA-core counts, board power (TGP), or prices have leaked at a credible grade for any of these cards. Figures for the 6070 Ti and 6060 tiers do not exist — only speculative codenames.

Where these specs come from (and why that matters)

The lineup rests on two separate leaker chains, and knowing which is which tells you how much to trust each part:

• Codenames + timing — kopite7kimi. The more credible leaker established that the gaming dies are the “GR20x” Rubin family and floated a second-half-2027 window (via WCCFtech, corroborated by VideoCardz).
• The spec numbers — Red Gaming Tech (RGT). Every VRAM, bus-width and performance figure traces to one March-2026 RGT video (written up by TweakTown, Overclock3D and others). TweakTown, OC3D and TechPowerUp re-reporting it is echo, not five independent leaks — and Tom's Guide argued the numbers don't add up.

The two chains agree on the big picture (Rubin, GR20x, TSMC 3 nm) but disagree on timing — which is exactly why the release date below is a debate, not a date.

The architecture: “Rubin” on 3 nm

The one claim with real corroboration is the family. The RTX 60-series is rumored to use NVIDIA's Rubin architecture — the same generation name NVIDIA uses for its next data-center AI GPUs — on a custom TSMC 3 nm process, with 6th-generation Tensor cores, 5th-generation RT cores, and a next DLSS 5 neural-rendering push. The through-line across leaks: gains are said to come from architecture (ray tracing, AI), not a big jump in raw core count. For reference, the rumored flagship's “up to 192 SM” is barely 13% more than the 5090's 170 — and that's the full die, so a shipping 6090 would be cut down from it.

One correction worth making, because several sites get it wrong in their comparison tables: the generation mapping is RTX 40 = Ada Lovelace, RTX 50 = Blackwell, RTX 60 = Rubin. If a “spec sheet” labels the 50-series as Ada, distrust the rest of it.

Release date: a genuine three-way split

There is no confirmed date, and the rumors don't converge — they fork:

The common thread pushing every estimate later is a global DRAM / GDDR7 shortage: NVIDIA is steering scarce memory toward far-more-profitable AI accelerators, showed no new GeForce at CES 2026, and even delayed the nearer-term RTX 50 Super refresh over the same 3 GB-module shortage (Digital Foundry roundup). A late-2027 launch would make the RTX 50→60 gap the longest in GeForce history.

Which card comes first — and when do the RTX 6070 and 6060 land?

Even once the series launches, the tiers don't arrive together. NVIDIA has staggered every recent generation: the flagship and high-end ship first, and the mid-range and budget cards follow months later. The RTX 50 series is the template — the RTX 5090 and 5080 launched in January 2025, the 5070-class followed in spring, and the 5060 tier came later still. Applying that cadence to the rumored RTX 60 windows:

NVIDIA has confirmed none of this — there is no per-tier date for any RTX 60 card. The ordering above is the historical GeForce cadence (RTX 40 and 50 both rolled out top-down) applied to the rumored series window, not a leaked schedule. If you specifically want a 6070 or 6060, plan for the back half of the rollout, whenever it starts.

What is not confirmed (the honest list)

• Any product from NVIDIA. No announcement, teaser, or official spec for a single RTX 60 card.
• All prices. No credible MSRP has leaked for any tier.
• CUDA-core counts, TGP. Not in any leak.
• The release date. Early 2027 vs H2 2027 vs 2028 remains unresolved.
• The 6070 Ti and 6060 tiers. No specs exist — only speculative codenames.
• The performance figures. ~2x path tracing and ~30-35% raster are pre-silicon targets from one leaker, not measured results.

Frequently asked questions

What GPUs are in the NVIDIA RTX 60 series?

Nothing is official. Leaks point to a Rubin-generation stack mirroring the RTX 50 lineup: an RTX 6090 flagship (GR202 die), an RTX 6080 (GR203), and an RTX 6070 (GR205), with lower tiers (6060 and below) existing only as speculative codenames. NVIDIA has not confirmed a single model, spec, or the lineup itself.

When will the RTX 60 series be released?

There is no confirmed date, and the rumors genuinely disagree. Three windows are in play: early 2027 (leaker Red Gaming Tech, widely doubted as too aggressive), the second half of 2027 (leaker kopite7kimi, the most-cited), and 2028 (a report from The Information, via Tom's Hardware). A global DRAM/GDDR7 shortage — NVIDIA steering memory to more-profitable AI chips — is the common reason every estimate keeps slipping later.

What architecture is the RTX 60 series?

“Rubin” — this is the one thing corroborated across leakers, and NVIDIA has confirmed Rubin as its next-generation architecture name (for data-center parts). The gaming dies are rumored to be the “GR20x” family on a custom TSMC 3 nm process. For the record, the correct generation mapping is RTX 40 = Ada Lovelace, RTX 50 = Blackwell, RTX 60 = Rubin — some sites' tables get this wrong.

How much VRAM will the RTX 6080 and 6070 have?

Rumored (single-source, provisional): 20 GB on the 6080 (320-bit bus) and 16 GB on the 6070 (256-bit) — both a step up from the 5080's 16 GB and 5070's 12 GB. These assume today's 2 GB GDDR7 modules; the leak itself says capacities could change with 3 GB modules. Nothing is confirmed.

Should I wait for the RTX 60 series or buy a GPU now?

If you need a card today, don't wait on a series rumored for 2027-2028 with no confirmed specs, prices, or dates. Buy from the current 2026 lineup. Waiting only makes sense if you already own an RTX 4090 or 5090 and care solely about maximum ray tracing — and even then the timeline is a rumor, not a plan.

How much will the RTX 60 series cost?

No price has leaked at any credible grade for any RTX 60 card. The only dollar figures in circulation are community guesses tied to the ongoing memory shortage. Treat every RTX 60 price you see as speculation, not a leak.

Is the RTX 60 series confirmed by NVIDIA?

No. NVIDIA has confirmed the Rubin architecture name but has shown zero consumer RTX 60 product, specifications, prices, or dates. Every number in circulation traces to leakers and projections.

When will the RTX 6070 be released?

There is no confirmed or leaked date for the RTX 6070 specifically. Even in the most optimistic rumor — a second-half-2027 series launch — the mid-range typically trails the flagship by months, so a 6070 realistically lands later in the window and slips toward 2028 on the more bearish reads. NVIDIA has confirmed nothing per-tier; treat any exact 6070 date as a guess.

When will the RTX 6060 be released?

No date has leaked for the RTX 6060 — the tier exists only as a speculative codename with no specs. Budget cards are historically the last to launch, several months after the flagship (the RTX 5060 followed the January-2025 RTX 5090 that May). So even if the series starts in late 2027, a 6060 realistically means 2028 or later. Any specific 6060 date you see is speculation.

When will the RTX 6080 be released?

Unconfirmed. The RTX 6080 (rumored GR203 die, ~20 GB GDDR7) would launch in the first wave alongside the 6090, so it tracks the series window itself: late 2027 at the earliest, 2028 on the bearish read. No separate 6080 date has leaked.

In the meantime: GPUs you can actually buy

A 2027-2028 rumor doesn't help you build a PC today. If you're shopping now, our best graphics cards of 2026 ranks the cards that are real, in stock, and sorted by value — and for AI and local-LLM builds specifically, our best GPUs for AI in 2026 weighs VRAM against price. Three worth a look across the range:

Curious how the rest of the next-gen field lines up? The flagship gets its own deep dive in our RTX 6090 specs & rumors page; AMD is rumored to answer with a high-end RDNA 5 flagship (Radeon RX 10900 XT) on a similar 2027 timeline; and the nearer-term RTX 50 Super refresh is the more immediate NVIDIA rumor. On the CPU side, the Zen 6 vs Nova Lake race is shaping up for the same CES 2027 window — all tracked here with the same honest framing. And if the real question is whether to wait at all, our 2026 GPU prices and the memory shortage explainer lays out the buy-now-vs-wait math.

Related buyer's guides

All reviews →

Buyer's guide / PC builds

Best gaming PC under $1,500 in 2026

The honest $1,500 build: a 16GB RX 9060 XT, Ryzen 5 and 32GB DDR5 for strong 1080p-ultra and entry-1440p gaming — with the real shortage-adjusted total, where every dollar goes, and whether a prebuilt beats it this year.

See the build →

Buyer's guide / PC builds

Best gaming PC under $2,000 in 2026

The value sweet spot: near-flagship 1440p gaming built around the RX 9070 XT and an X3D CPU — with the honest 2026 total, budget allocation, and exactly how to duck under a hard $2,000.

See the build →

Buyer's guide / PC builds

Best gaming PC under $3,000 in 2026

Where 4K becomes real: the fastest gaming CPU paired with an RTX 5080 for 1440p-max and 4K-high with DLSS 4 — the honest shortage-adjusted total, where every dollar goes, and a matched monitor.

See the build →

Buyer's guide / PC builds

Best gaming PC under $5,000 in 2026

The honest high-end build: a fully-maxed RTX 5080 creator-gaming machine with a 16-core 9950X3D, 64GB and 4TB — plus the truth about why $5,000 does not reliably buy a 5090 at 2026 street prices.

See the build →

Buyer's guide / AI PCs

LLM VRAM calculator

Enter a model size, quantization and context length to estimate the VRAM or unified memory an LLM needs to run locally — and instantly see which 2026 machines can hold it.

Use the calculator →

Buyer's guide / AI PCs

How to run LLMs locally: the 2026 hardware guide

How much VRAM or unified memory you need to run 7B-120B models locally, the capacity-vs-bandwidth tradeoff, and the 2026 hardware that hits each point on the curve — GB10, Ryzen AI Max, Apple Silicon and RTX GPUs.

Read the guide →