Illustration: a concept graphics card with next-generation memory modules. This is an AI-generated illustration, not a photograph of any real NVIDIA product.
The Short Version: NVIDIA's RTX 50 Super Refresh Is Reportedly Slipping Into 2027
If you have been waiting to buy a graphics card until NVIDIA launches its RTX 50 Super refresh, the latest reporting has bad news: the lineup that was expected to land in 2026 is now rumoured to arrive at CES 2027 in January, with retail availability following in the first half of 2027. The reason is not a lack of finished silicon — it is a lack of memory. According to supply-chain reporting, NVIDIA's board partners already have completed Super GPUs in hand, but the launch is being held back because the 3GB GDDR7 memory modules these cards need are scarce, with memory makers prioritising AI data centre demand.
Here is the critical honesty clause for this entire article: NVIDIA has never officially announced the RTX 50 Super series. Every date, every specification and every price in the leak record below is a rumour — reported by outlets like VideoCardz, TweakTown, Tech4Gamers and Tech Insider, citing leakers such as BenchLife and MEGAsizeGPU. Wherever a fact is confirmed, this article says so; wherever it is only reported, it is labelled as a report. That distinction matters more than ever in late 2026, because conflicting leaks disagree about whether any Super card could still appear in 2026.
TL;DR — The Key Facts in 60 Seconds
- The rumoured lineup: four cards — RTX 5060 Super (12GB), RTX 5070 Super (18GB), RTX 5070 Ti Super (24GB) and RTX 5080 Super (24GB). All rumoured; nothing confirmed by NVIDIA.
- The rumoured launch: reveal at CES 2027 in January, on shelves in Q1–Q2 2027. First reported by BenchLife via VideoCardz and echoed by TweakTown, Tech4Gamers and Tech Insider across June–August 2026.
- The reason for the delay: a shortage of high-density 3GB GDDR7 modules, which memory suppliers are directing toward AI accelerators instead of gaming cards.
- The point of the refresh: VRAM, not speed. Leaks consistently say the Supers keep the Blackwell architecture and add memory — 18GB and 24GB where the current cards have 12GB and 16GB.
- What comes after: the RTX 60 series (reportedly "Rubin") is rumoured for 2028–2029, meaning the Super refresh would be the only new NVIDIA gaming lineup for a long stretch.
- What to do now: if your current GPU still runs your games, waiting costs you nothing — but waiting until 2027 for an unpriced, unconfirmed card is a gamble; this article gives you the full buy-or-wait framework.
Confirmed vs Rumoured: What Is Actually Known
Before going further, here is the honest scorecard — confirmed facts separated from reported claims, as of 31 October 2026.
| Status | Claim | Source |
| CONFIRMED | NVIDIA has made zero official announcements about any RTX 50 Super card | Tracked by FlowUp and Tech Insider through September 2026 |
| CONFIRMED | 3GB GDDR7 modules exist and are in short supply, with AI demand taking priority | Supply-chain reporting; NVIDIA told Tom's Hardware memory supply is constrained |
| CONFIRMED | NVIDIA released no new consumer GeForce card in 2026 | Industry reporting; first such year in three decades |
| CONFIRMED | RTX 5090 has not been discontinued and keeps shipping | NVIDIA SVP Jeff Fisher, per Madshrimps |
| RUMOUR | RTX 50 Super reveal at CES 2027, retail Q1–Q2 2027 | BenchLife via VideoCardz; TweakTown; Tech4Gamers; Tech Insider |
| RUMOUR | Lineup: 5060 Super 12GB, 5070 Super 18GB, 5070 Ti Super 24GB, 5080 Super 24GB | Multiple leakers, consistent across outlets |
| RUMOUR | Board partners already hold finished Super silicon; NVIDIA HQ is holding the launch | Hardvance, via Tech Insider |
| RUMOUR | RTX 60 series ("Rubin") delayed to 2028–2029 | Moore's Law Is Dead and others |
| DISPUTED | Some leaks claim a 2026 Super launch is still possible | Noted by Tom's Hardware, via TechTimes |
The main point: everything about the Super refresh's existence, date and specifications is unconfirmed. Treat every "will launch" sentence in this article — and anywhere else — as "is reported to launch" until NVIDIA speaks.
The Rumoured Lineup: Four Cards, One Theme — More VRAM
Despite the date disputes, the reported specifications have stayed remarkably consistent across every outlet since mid-2026. That consistency is the strongest signal in the whole leak record: when TweakTown, Tech4Gamers, Tech Insider and TechTimes all describe the same four cards with the same memory figures, the underlying source material is probably the same set of board-partner documents. Here is the rumoured lineup:
| Card (rumoured) | Rumoured VRAM | Current card it refreshes | Current VRAM |
| RTX 5060 Super | 12GB GDDR7 | RTX 5060 (8GB) | 8GB |
| RTX 5070 Super | 18GB GDDR7 | RTX 5070 (12GB) | 12GB |
| RTX 5070 Ti Super | 24GB GDDR7 | RTX 5070 Ti (16GB) | 16GB |
| RTX 5080 Super | 24GB GDDR7 | RTX 5080 (16GB) | 16GB |
The pattern is unmistakable: this refresh is about memory capacity, not raw speed. Leaks describe only modest core-count bumps alongside the big VRAM jumps — the same playbook NVIDIA used for the RTX 40 Super series, which was also unveiled at CES (in 2024). The 3GB GDDR7 modules are the enabler: they let NVIDIA raise capacity without redesigning the memory bus, so a 256-bit card can carry 24GB instead of 16GB simply by swapping denser chips onto the same board design.
The standout of the rumoured set is the RTX 5060 Super with 12GB. The current RTX 5060's 8GB frame buffer has been its most criticised specification, and a 12GB variant would directly answer the loudest complaint about the entry-level Blackwell lineup — a meaningful upgrade for 1080p and 1440p gamers, assuming the price stays near the current card's territory. Nothing about pricing has leaked credibly, so treat any rumoured price as fiction until an official MSRP appears.
Why the Delay? The 3GB GDDR7 Bottleneck
To understand why finished graphics cards would sit in warehouses instead of on shelves, you have to understand what the Super refresh is made of. The entire rumoured lineup depends on high-density 3GB GDDR7 memory modules — chips that pack 3 gigabytes into a single package, versus the 2GB modules used on the current RTX 50 series. Swap 2GB chips for 3GB chips on the same board, and a 16GB card becomes a 24GB card with no other changes. It is an elegant, cheap refresh strategy — but only if you can actually buy the chips.
Right now, NVIDIA apparently cannot buy enough of them. The reporting converges on a single cause: the AI boom is eating the memory supply. The same GDDR7 modules — and the HBM memory adjacent to them — are needed by NVIDIA's AI data centre accelerators, which sell for tens of thousands of dollars each and carry far higher margins than any gaming card. Memory makers are rational businesses; when an AI customer will take every module you can produce, the gaming division waits in line. TweakTown's June 2026 report put it bluntly: if the AI-driven memory crisis did not exist, the Super series would have released earlier this year.
The most vivid detail in the leak record comes from supply-chain reporting via Hardvance, summarised by Tech Insider: NVIDIA's board partners already have finished Super silicon in hand, while NVIDIA's head office is holding the launch back over memory supply. If true, that means the delay is a deliberate allocation decision, not a technical problem — the cards exist, the chips to complete them in volume do not, and NVIDIA would rather wait than launch a paper refresh it cannot supply. That is also consistent with what NVIDIA told Tom's Hardware: demand for GeForce RTX is strong, but memory supply is constrained.
The main point: the RTX 50 Super delay is a memory-supply story, not a GPU-design story. The same shortage is pushing up DRAM contract prices across the industry — TrendForce's latest forecast has conventional DRAM rising another 10 to 15 percent in the fourth quarter of 2026, with NAND flash up 15 to 20 percent — which is why graphics card street prices have been climbing all year instead of falling.
How the Delay Story Unfolded: A Timeline of Reports
The Super delay did not arrive as a single announcement — it accreted over months of leaks. Here is the timeline as reported, which also shows why the CES 2027 date carries weight: it is not one outlet's guess, it is the same conclusion reached independently across three separate reporting windows.
- June 2026 — TweakTown: the first major report that the Super series is "now rumored to be coming in early 2027," explicitly blaming the AI-driven memory crisis and noting the lineup was never formally announced, so every date is technically rumour.
- June–July 2026 — Tech Insider tracking: independent tracking concluded the RTX 5080 Super's release looked delayed to CES 2027 rather than shipping at all in 2026, citing zero official NVIDIA announcements across the entire Super lineup.
- Summer 2026 — MEGAsizeGPU: the well-known leaker reported the Super series was "back in development," possibly including the 12GB RTX 5060 — suggesting the lineup had been paused and restarted rather than cancelled.
- August 2026 — BenchLife via VideoCardz: the Taiwanese outlet's report, amplified by VideoCardz, set the most specific target yet — an announcement at CES 2027 in Las Vegas with launch soon after — and identified 3GB GDDR7 availability as the primary gating factor. Notebookcheck and PCGuide echoed the same window.
- 31 August 2026 — Tech Insider follow-up: a check confirmed that as of that date there was still no official date, price, or acknowledgment for the 5080 Super, 5070 Ti Super or 5070 Super.
- October 2026 — the dispute: TechTimes summarised the honest state of play — most outlets point to CES 2027, but Tom's Hardware noted some leaks still claim a 2026 launch is possible. NVIDIA's silence means neither side can be proven right yet.
The main point: the CES 2027 window is the consensus of the leak record, not a confirmed date. The one dissenting thread — a possible 2026 launch — is worth remembering precisely because NVIDIA has said nothing; a surprise late-2026 paper launch of one card cannot be ruled out, even if the weight of reporting sits firmly in January.
What CES 2027 Could Look Like
CES in Las Vegas has become NVIDIA's traditional stage for mid-cycle GeForce refreshes — the RTX 40 Super series was unveiled at CES 2024, so a CES 2027 Super reveal would follow an established pattern. If the reporting holds, expect a keynote announcement in the first week of January 2027 with retail availability staggered through Q1 and into Q2 2027.
What would actually be announced? Based on the consistent leak record: the four cards above, positioned as VRAM upgrades on the Blackwell architecture — no new architecture, no dramatic core-count leaps, and pricing that NVIDIA will set against a market where current cards are selling far above their launch MSRPs. That pricing decision is the great unknown. If NVIDIA prices the 5070 Super near the current 5070's $549 launch MSRP while street prices sit near $800, the Super cards would look like bargains on day one. If instead NVIDIA raises MSRPs to match the inflated market — as some fear — the refresh would simply ratify the new, higher price floor.
The main point: CES 2027 is the rumoured stage, but the rumoured script has no confirmed prices. Any article that tells you what a 5080 Super will cost is inventing numbers — this one will not.
Illustration: graphics memory modules under inspection. This is an AI-generated illustration, not a photograph of any real hardware.
The VRAM Story: Why 12GB, 18GB and 24GB Are the Whole Point
The rumoured Super refresh is, at its core, an admission: the current RTX 50 series does not have enough video memory for the prices it commands. A $549 RTX 5070 with 12GB and a $999 RTX 5080 with 16GB looked thin on VRAM at launch in early 2025, and they look thinner now that games and AI workloads keep demanding more. The Super cards would fix exactly that — 18GB on the 5070 Super, 24GB on the 5070 Ti Super and 5080 Super — using denser 3GB GDDR7 modules on the existing board designs.
Why does this matter so much? Three reasons. First, modern games at 1440p and 4K with high-resolution texture packs can already exceed 12GB in the most demanding titles, and the trend points only upward. Second, local AI workloads — image generation, LLMs, video tools — are VRAM-hungry, and 24GB turns a gaming card into a genuinely capable AI workstation card. Third, VRAM is the one specification that cannot be upgraded later: you can overclock a GPU, but you cannot add memory to it. Buying a card with too little VRAM is the most permanent mistake in PC building, which is why the Super refresh's memory-first approach resonates with buyers even though the architecture is unchanged.
The main point: if the leaks are right, the Super series is the VRAM correction the RTX 50 lineup needed from day one — which is also why the delay stings. The cards buyers most want to wait for are the cards that need the scarcest components.
What Comes After: RTX 60 Delayed to 2028–2029
The Super delay looks different once you see what sits behind it. According to industry insiders including Moore's Law Is Dead, NVIDIA's next full GeForce generation — the RTX 60 series, reportedly based on the "Rubin" architecture — has slipped to 2028–2029, with its exact architectural direction not yet finalised. If that reporting is right, the RTX 50 Super refresh is not a quick bridge to something newer — it is the only new NVIDIA gaming lineup for the next two-plus years, a mid-cycle stepping stone sandwiched between the 2025 Blackwell launch and a 2028 architecture change.
This also reframes NVIDIA's 2026. Per the same insider reporting, NVIDIA confirmed it would release no new consumer GeForce card in 2026 — the first such year in thirty years. The company's public messaging at Computex 2026 centred instead on its RTX Spark chips for AI PCs. Gaming, the business that built NVIDIA's brand, spent 2026 in a holding pattern while the data centre business consumed the memory supply. For gamers, the practical consequence is stark: the card you buy today may need to last until 2028, because the next true generational leap is years away.
The main point: the Super refresh matters more than a normal mid-cycle update, because nothing newer is coming behind it for a long time. That raises the stakes of the buy-or-wait decision at the end of this article.
AMD's Side: RDNA 5 Waits Too
NVIDIA is not alone in the waiting room. AMD's next GPU architecture, RDNA 5, is not expected before late 2027 or even 2028, with only one AMD chip (reportedly "AT2") rumoured for 2027 at all. Neither camp is in a rush, because neither camp faces pressure from the other — and both face the same memory shortage. For the full RDNA 5 picture, see our earlier deep dive: AMD RDNA 5 Graphics Cards Leaked: Everything We Know About the Radeon RX 10000 Series (2027–2028).
What AMD did release in 2026 tells the same thin-year story as NVIDIA's silence: the Radeon RX 9070 GRE — essentially a revised GPU with slightly reduced specifications rather than a new class of hardware — and that was the year's headline AMD launch. The competitive dynamic this creates is unusual: with both vendors' next architectures pushed out, the current lineups have to carry the market through 2027, which is part of why street prices have stayed high. There is no new-generation price pressure forcing old cards down.
The main point: the GPU market's 2027 depends on refreshes, not revolutions — Super cards from NVIDIA, and not much new from AMD. Buyers hoping competition will collapse prices are likely to be disappointed.
2026: The Year the GPU Market Stood Still
Step back and 2026 looks remarkable for what did not happen. One industry summary put it plainly: the year is being filled with small revisions, delayed launches and product refreshes that feel more incremental than transformational. The complete list of genuinely new 2026 desktop GPU launches from the two big vendors is essentially: the RX 9070 GRE (a cut-down revision), an RTX 5070 12GB laptop variant (more memory on an existing chip), and a rumoured RTX 5050 9GB (a tweaked existing GPU, not an architectural leap). That is the whole year.
The used market reflects the stagnation. The RTX 4090 — launched in October 2022 — remains the second-fastest gaming GPU you can buy, and at its launch MSRP it could plausibly stay near the top of the charts for six years or more before a true successor surpasses it. When a four-year-old card is still the card to beat, generational progress has stalled — and the stall is structural, driven by memory supply and AI demand, not by any single company's roadmap choices.
The main point: 2026 may go down as one of the weakest years for new GPU launches in the industry's history — which is exactly why the CES 2027 Super reveal, if it happens, will land with unusual force.
GPU Prices Right Now: The October 2026 Reality Check
Waiting for CES 2027 only makes sense if you know what buying today actually costs. The answer, across every price tracker in October 2026, is: far above launch MSRP, and climbing. Tom's Hardware's GPU price tracker on 8 October 2026 showed the damage clearly:
| Card | Launch MSRP | Lowest US price, 8 Oct 2026 | Premium over MSRP |
| RTX 5070 | $549 | $799 | +46% |
| RTX 5070 Ti | $749 | $1,179 | +57% |
| RTX 5080 | $999 | $1,599 | +60% |
| RX 9070 XT | $599 | $831 | +39% |
And prices kept rising through October. An Amazon snapshot on 30 October 2026 found the cheapest new RTX 5070 at $960.99 (75% over MSRP, from a marketplace seller), the RTX 5070 Ti at $1,466.69 (96% over), and the RTX 5080 at $2,211.17 (121% over). The XFX Swift RX 9070 XT was $919.99 — 54% over its $599 MSRP, but $546.70 cheaper than its direct CUDA rival, the cheapest RTX 5070 Ti. Two bright spots in the gloom: the RX 9070 GRE 12GB was the only current card selling under MSRP ($529.99 against a $549 launch price on 6 October), and the Intel Arc B570 10GB remained the cheapest card near MSRP, with the RX 9060 XT 16GB the cheapest 16GB card listed. For the full October price landscape, see Graphics Card Prices in October 2026: Why GPUs Cost So Much.
The main point: no current card is within 10% of its MSRP, and the median price rise across current cards was about 13% in October alone. This is the market the Super refresh would launch into — and the market you are buying in if you buy today.
The RTX 5090 Question: Not Discontinued, Still Absurdly Priced
One rumour deserves a direct answer because it keeps resurfacing: no, the RTX 5090 has not been discontinued. NVIDIA SVP Jeff Fisher has said the flagship keeps shipping, and some leakers' claims otherwise are contradicted by the company's own statements. What is true is that the 5090 is priced like a scarce card — roughly 3.6 times its $1,999 MSRP in October snapshots, reselling in the $5,000–$6,000 range — because it is effectively a scarce card. A formal end-of-life notice, if one ever comes, would mostly confirm what the market already knows.
The more interesting related rumour: some leakers claim the RTX 5080 16GB could be discontinued in favour of the 24GB Super version in early 2027. Nothing about this is official, but the logic is sound — a direct replacement is the classic trigger for current-model prices to soften. Our full analysis of the flagship's strange 2026 is here: RTX 5090 Shortage Explained (October 2026).
The main point: the 5090 is not going anywhere officially, but its street price says scarcity — and a 5080 Super could eventually replace the 16GB 5080.
Buy Now or Wait for the Super Refresh? The Honest Framework
This is the question every section of this article has been building toward. Here is the framework, with no hedging:
Buy now if: your current GPU cannot run the games or workloads you need today; you find a card at or near its launch MSRP (the RX 9070 GRE under MSRP and the Arc B570 near MSRP are the current examples); or your purchase is for work that earns money now — a creator or developer losing billable hours to an old card should not wait six months for a rumour.
Wait for CES 2027 if: your current card still plays your games acceptably; you specifically want 18GB+ VRAM for 4K longevity or local AI work; or you are buying at the $700+ tier, where the rumoured 5070 Ti Super and 5080 Super's 24GB would be the most meaningful upgrades over today's 16GB cards.
The case against waiting: the Super lineup is unconfirmed, its prices are unknown, its retail availability could slip into mid-2027, and launch-day stock in a memory crisis could be thin — meaning real availability at sane prices might be a year away. Waiting for a rumour means paying today's inflated prices later anyway if the rumour disappoints.
The case for waiting: if the leaks are right, the 5070 Super's 18GB and the 24GB Supers fix the single biggest weakness of the current lineup, and buying a 12GB or 16GB card weeks before a VRAM-doubled refresh launches is the classic buyer's regret scenario.
The main point: waiting is rational only if your current GPU is tolerable — otherwise you are trading six months of actual use for a rumour with no confirmed price.
What to Buy Right Now: The Best Pick at Every Budget
If the framework above says "buy now," here is the October 2026 shortlist — every pick grounded in the price snapshots cited above, not in wishful thinking:
- Under $350 — Intel Arc B570 10GB: the cheapest card anywhere near MSRP. Not the fastest, but the only sane entry point while everything else is inflated. Check driver maturity for your specific games first.
- $500–$650 — RX 9070 GRE 12GB: the only current-generation card that has sold under its MSRP. 12GB is workable at 1440p, and at this price it undercuts the RTX 5070's street price by hundreds.
- $700–$950 — RX 9070 XT 16GB: the high-end pick on value. 16GB on a 256-bit bus, and in October it cost over $500 less than its CUDA rival, the RTX 5070 Ti. The catch, as ever: skip it if your software needs CUDA — for Blender, DaVinci Resolve's AI tools, or local LLM work, NVIDIA's ecosystem still wins.
- $950–$1,200 — RTX 5070 Ti 16GB (if found near the low end): the rational NVIDIA pick for 4K high-refresh gaming and AI work, but only if you can find one near the $1,179 October floor rather than the $1,466 marketplace extremes. Above that, walk away.
- Over $1,500 — wait: paying $1,599+ for an RTX 5080 or $5,000+ for a 5090 in late 2026, months before a rumoured 24GB refresh, is the worst value proposition on this list. This is the tier where waiting costs least and gains most.
The main point: the value winners in October 2026 are AMD's 9070-series and Intel's B570 — NVIDIA's cards carry the biggest premiums over MSRP.
The Used-Market Angle: Sell High, Buy Smart
The inflated new-card market has an odd side effect: used cards are holding their value, because new ones cost so much. That makes the upgrade math unusual — selling the card you are replacing is currently the most reliable way to lower the net cost of an upgrade. A used RTX 3080 or 4070 still commands strong money, which can fund most of a 9070 XT purchase.
Two cautions. First, an online quote is an estimate until the card is tested — verify any used purchase in person where possible. Second, the rumoured 5080 Super is the classic trigger for used-flagship prices to drop: if a 24GB 5080-class card launches at CES 2027, used RTX 4080 and 5080 prices should soften after. Buyers hunting the used market may want to time purchases for just after the reveal, not before.
The main point: in this market, your old GPU is an asset — sell it strategically and the upgrade costs far less than the sticker prices suggest.
Black Friday 2026: What to Realistically Expect for GPUs
Black Friday falls on 27 November 2026, with Cyber Monday on the 30th — about four weeks away. Here is the honest outlook, because the usual "doorbuster GPU deals" narrative does not apply this year:
- Expect $30–$100 off inflated street prices — not a return to launch pricing. Most projections point to modest cuts, and a 20% discount on an RTX 5080 at today's prices would still land above where the card sat in August.
- A card at or near its launch MSRP is the real win. Last Black Friday, the RTX 5070 sold for $479–$499 and the RX 9070 XT hit its $599 list price; those prices are the best case this year, not the expectation.
- Do not trust the discount badge. Write down the October street price of the exact model you want today, and judge any Black Friday "deal" against that number — a big percentage off an inflated "was" price can still leave you paying more than the card cost in summer.
- Early deals start in the first two weeks of November, following last year's pattern — you do not have to wait for the Friday itself.
The main point: treat Black Friday 2026 as a chance to buy at October prices with a small sweetener — not as the day GPUs get cheap again.
7 Mistakes to Avoid in the Super-Delay Era
- Treating rumoured specs as confirmed. The four-card lineup is consistent across outlets but still unannounced. Do not plan a build around a card that may not exist.
- Paying marketplace-seller premiums. October snapshots show marketplace sellers charging 75–121% over MSRP. If the listing is not sold by the retailer or the card maker, the price is not the price.
- Buying 8GB cards at 2026 prices. With a rumoured 12GB 5060 Super on the horizon and 8GB already limiting, paying inflated prices for 8GB is the worst longevity bet.
- Ignoring VRAM for AI work. If you do local image generation or LLMs, 12GB is the floor and 16GB+ is the target — the Super refresh exists precisely because buyers learned this the hard way.
- Waiting without a working GPU. Six months of no gaming or no work output is a real cost. A rumour is not worth a dead PC.
- Forgetting the power supply. AMD rates the 9070 XT board at 304W — a Super-class upgrade may need PSU headroom you do not have. Budget the PSU, not just the card.
- Assuming CES means available. A January reveal with Q1–Q2 retail means real stock at sane prices could be a spring or summer story. "Announced" and "buyable" are different words.
Frequently Asked Questions
1. Is the RTX 50 Super series confirmed?
No. NVIDIA has never officially announced it. Everything — the four cards, the CES 2027 date, the VRAM figures — is reported by leakers and press, not confirmed by NVIDIA.
2. When will the RTX 50 Super cards launch?
The consensus of the leak record points to a reveal at CES 2027 in January, with retail availability in Q1–Q2 2027. A minority of leaks suggest a 2026 launch is still possible, but most reporting disagrees.
3. Which cards are rumoured?
RTX 5060 Super (12GB), RTX 5070 Super (18GB), RTX 5070 Ti Super (24GB) and RTX 5080 Super (24GB) — all GDDR7, all on the Blackwell architecture.
4. Why was the Super refresh delayed?
According to supply-chain reporting, the 3GB GDDR7 modules the cards need are in short supply because memory makers are prioritising AI data centre demand. Board partners reportedly already hold finished Super silicon.
5. Will the Super cards be faster than the current ones?
Leaks describe the refresh as a VRAM upgrade with only modest core-count bumps — expect meaningfully more memory, not a generational speed leap.
6. What will the RTX 50 Super cards cost?
Unknown. No credible pricing has leaked. Any price you see quoted before NVIDIA announces is speculation.
7. Is the RTX 5090 discontinued?
No. NVIDIA SVP Jeff Fisher has said the 5090 keeps shipping. Its extreme street price reflects scarcity, not discontinuation.
8. When is the RTX 60 series coming?
Rumoured for 2028–2029 ("Rubin" architecture) — meaning the Super refresh would be the only new NVIDIA gaming lineup for years.
9. Should I buy an RTX 5070 now or wait for the 5070 Super?
If your current GPU is unusable, buy now — preferably the best value at your budget. If it still runs your games and you want 18GB VRAM, waiting for CES 2027 is rational, but remember the card is unconfirmed and unpriced.
10. Will Black Friday 2026 have good GPU deals?
Expect small discounts ($30–$100) off already-inflated street prices. A card at or near its launch MSRP would be the real win; do not expect a return to 2025 pricing.
Conclusion: The Waiting Game, Played Honestly
The RTX 50 Super story is really two stories. The first is a product story: four rumoured graphics cards that would fix the RTX 50 series' biggest weakness — not enough VRAM — using denser 3GB GDDR7 modules, most likely revealed at CES 2027. The second is a market story: an AI-driven memory shortage that has inflated GPU prices all year, frozen both vendors' roadmaps, and turned 2026 into one of the quietest GPU years in memory.
For buyers, the honest summary fits in three sentences. If your GPU works, waiting for January costs you nothing and might gain you 18–24GB cards. If it does not work, buy the best value available today — the 9070 GRE, the 9070 XT, or the Arc B570 — and do not pay marketplace-seller premiums for the privilege. And whatever you do, treat every Super specification and date as a rumour until NVIDIA itself speaks, because in this market, the only thing more expensive than an overpriced GPU is a purchasing decision built on a leak.
This is a developing story. If NVIDIA announces the Super lineup — or if the CES 2027 reporting shifts — this guide will be updated with confirmed details.
Sources
Related reading on MyThrift Hub: AMD RDNA 5 / RX 10000 Series: Everything We Know · RTX 5090 Shortage Explained (October 2026) · Graphics Card Prices in October 2026
All images in this article are AI-generated illustrations, not photographs of real products. Prices cited are US street-price snapshots dated October 2026 and will change; rumoured specifications are marked as rumours throughout.
Comments
Post a Comment