NVIDIA RTX 60 Series Reportedly Delayed to 2028: A PC-Building Winter as Gaming GPUs Face Marginalization#

Zixing | September 22, 2026

Hi everyone, I’m Zixing.

This week the hardware world was hit by a bombshell, with all the talk focused on next-generation gaming GPUs. Leaker Kopite7Kimi reported that NVIDIA’s next-generation GPUs based on the Rubin architecture (internal codename GR20x, corresponding to the RTX 6080/6200) will be pushed back to 2028, and that the originally planned Super series refresh may be cancelled along with it. Add to that NVIDIA’s reclassification of GeForce under “edge computing” in its financial filings, the absence of any separate gaming mention in fiscal Q2 2027, and the reality of the RTX 5090 soaring from its $1,999 MSRP to around $9,000 — and the phrase “the marginalization of gaming GPUs” in the headline is not clickbait, but a business logic that can be traced out step by step.

First, the conclusion: within NVIDIA’s strategic map, gaming GPUs have already been demoted from “growth engine” to “something done on the side”. This is not product-level laziness, but the compounded result of a triple pivot in financial framing, capacity allocation, and pricing structure. Whether to wait or buy depends on first taking this logic apart.

Table of Contents#

1. The Leak and Timeline: RTX 60 Delay and Super Cancellation
2. Financial Review: Why GeForce Was Filed Under “Edge Computing”
3. The Rubin Delay’s Chain Effects on Supply and Pricing
4. The RTX 5090’s Pricing Logic from $1,999 to $9,000
5. The “VRAM Crisis” at Month 10: The Real Situation
6. Where PC Builders and the Wait-and-See Crowd Actually Stand
7. When to Buy a GPU: Zixing’s Practical Advice

1. The Leak and Timeline: RTX 60 Delay and Super Cancellation#

First, put Kopite7Kimi’s leak on the table. This leaker’s hit rate on NVIDIA GPU news has long been fairly high, and the information volume this time is substantial.

1.1 New-Generation GPU Naming and Codenames#

1.2 What Recent Generations Say About the Slowing Cadence#

Stretch the timeline out, and you can clearly see the update cadence of NVIDIA gaming GPUs being stretched — even scrambled.

One trend stands out from this table: the old rhythm of a refresh every two years, bridged by Super models in between, is being broken. If the Super line is indeed cancelled, then between the RTX 50 and the RTX 60, players face a long gap window with little mid-cycle reinforcement.

1.3 The Leak’s Credibility and Open Questions#

• Supporting factors: Kopite7Kimi’s prior leaks on the Blackwell architecture and RTX 50 series specs were largely on target, and the Rubin and GR20x codename details make this round of news look well-grounded.
• Open questions: a one-year delay is a big move for a company of NVIDIA’s scale, and it has never been officially confirmed; the leaker’s statement is also not a final confirmation, and could still be revised with changes in capacity and market demand.
• Objective judgment: even if the 2028 date proves fuzzy, the broader direction of a slowing gaming-GPU cadence can already be cross-verified from the financials and pricing — expanded below.

2. Financial Review: Why GeForce Was Filed Under “Edge Computing”#

This is the layer most worth digging into. NVIDIA is not saying “gaming doesn’t matter” with words — it has written it into the institutional record through its financial reporting.

2.1 The Quiet Change in Financial Classification#

NVIDIA has long disclosed data by business line. Gaming used to be an independent and prominent line, broken out within the product categories. But in the most recent filings, two notable changes appeared:

2.2 The Revenue-Structure Shift Is the Root of “Marginalization”#

Financial classifications are never adjusted out of thin air — behind them lies a real change in revenue structure. A rough accounting makes the direction clear.

• When compute chips’ gross margins, revenue scale, and growth slope all far exceed gaming GPUs, product planning, wafer capacity allocation, and launch cadence naturally tilt toward the data center.
• Gaming GPUs are not “disliked” — they are simply queued behind in resource priority. The most visible manifestation is wafer capacity allocation: AI chips, with their far higher per-die value, command stronger negotiating power on TSMC’s advanced nodes.

2.3 What the “Edge Computing” Label Really Means#

Filing GeForce under “edge computing” is technically defensible — gaming terminals, laptops, and handhelds do carry some edge-side compute. But the classification also cleverly downgrades gaming GPUs’ status as an independent category, robbing them of standalone visibility in the financial presentation. To capital markets, the signal is unmistakable: NVIDIA no longer wants gaming GPUs to carry its growth expectations.

3. The Rubin Delay’s Chain Effects on Supply and Pricing#

A delayed product launch is never just “coming out later” — it triggers a chain reaction on both the supply side and the pricing side.

3.1 Direct Supply-Side Effects of the Delay#

3.2 The Pricing Logic#

A delayed launch means existing products must carry the high-end segment for longer, and vendors then have no incentive to cut prices.

• No supply increase: with the RTX 60 series absent, no new high-end supply arrives to dilute prices, and flagships like the RTX 5090 have that much more footing to hold or raise prices.
• No price anchor: the Super series would normally provide an anchor of “more performance at the same price” or “relatively affordable”; once cancelled, that premium-suppression mechanism is gone.
• Capacity siphoned by AI: even in 2028, Rubin output will first guarantee data-center demand, with the share flowing to the gaming market uncertain.

3.3 A Reference Table for “Will Prices Fall?”#

4. The RTX 5090’s Pricing Logic from $1,999 to $9,000#

This is the single most illustrative case. A flagship card whose price can multiply within about a year over its MSRP reflects the triple stack of “marginalization + tight supply + rigid demand.”

4.1 Price Evolution Timeline#

4.2 How It Got This Far#

• AI demand spillover: individuals and studios doing AI inference and small-scale training found consumer flagship cards’ VRAM and compute “good enough,” and began competing with gamers for the same cards — driving prices straight up.
• No new product in sight: the Rubin delay plus the Super cancellation removed the expectation of “waiting for the next generation means lower prices,” so holders and channels grew bolder in hoarding and quoting.
• Flagship scarcity: as the top consumer card, the RTX 5090’s output is inherently limited; under supply-demand imbalance, price is almost entirely seller-dictated.

4.3 What It Really Means#

We all understand: at this price point, no one is paying “for gaming” — they are paying “for compute.” When a gaming GPU’s price is hostage to AI demand, it shows this product line has left pure gaming-market logic and entered the compute market — the most ironic landing of “marginalization.”

5. The “VRAM Crisis” at Month 10: The Real Situation#

Corresponding to flagship cards being speculated into the thousands of dollars is the “VRAM crisis” ordinary players care about most. Now in its tenth month, here is its real current state.

5.1 Where the VRAM Crisis Came From#

Gaming VRAM (GDDR7, GDDR6X) faces a supply squeeze similar to server HBM: packaging, wafer, and memory die capacity is being heavily absorbed by AI demand, pushing consumer memory die quotes steadily higher.

5.2 The Crisis’s Actual Impact on Build Costs#

This crisis shows up not only in flagship cards — mid-range GPU and memory price hikes hit ordinary players more directly.

5.3 What Month 10 Means#

• This is no longer “short-term shortage” volatility, but structural supply tightness lasting nearly a year.
• It shares one logical chain with the Rubin delay and GeForce’s marginalization: AI compute demand is restructuring capacity allocation across the entire semiconductor supply chain, and gaming GPUs sit relatively far back in resource priority.
• For players, the most realistic problem: the “VRAM crisis” shows no systemic inflection point toward relief in the short term, because the AI demand behind it is long-cycle.

6. Where PC Builders and the Wait-and-See Crowd Actually Stand#

Stacking the layers above and returning to people — where exactly do PC builders and the wait-and-see crowd stand right now?

6.1 The Wait-and-See Crowd’s Dilemma#

The “wait-and-see crowd’s” default logic of “wait a while and prices fall naturally” has almost entirely failed in this market:

• No new product in sight: waiting for the RTX 60 series may well mean waiting until 2028, with price and performance expectations full of uncertainty.
• No Super hope: waiting for the mid-cycle refresh could mean waiting for nothing at all.
• Price drops unlikely: with long-tight supply, no new launches, and AI demand siphoning continuously, both high-end and mid-range cards lack external pressure for major price cuts.

6.2 PC Builders’ Actual Situation#

6.3 Conclusion: The Nature of This Cycle Has Changed#

Past price hikes were cyclical — wait long enough and they would fall; but this round of hikes is structural — behind it lies AI compute’s restructuring of supply-chain capacity, plus NVIDIA’s repositioning of gaming GPUs strategically. Applying the wait-and-see crowd’s old experience to the current market most likely yields “more expensive, longer, fewer choices.”

7. When to Buy a GPU: Zixing’s Practical Advice#

Finally, as a hardware editor, combining this round’s specific conditions of “marginalization + delay + VRAM crisis,” here are a few grounded, slogan-free recommendations.

7.1 First Ask Yourself: “Must-Have” or “Can Wait”#

7.2 The Practical Priority Order for a Budget#

In a rising market, spending the limited budget on the scarcest, least-likely-to-drop components is the most pragmatic choice.

7.3 Set a Dual Stop-Loss: Time and Budget#

Waiting should not be aimless. I recommend setting clear boundaries on “waiting”:

• Time stop-loss: tell yourself exactly how long you’ll wait. For example, “wait until mid-2027; if there’s still no solid news on new products, give up waiting” — avoiding being dragged indefinitely by uncertain leaks.
• Budget stop-loss: set an acceptable ceiling. In real conditions, many wait-and-see players don’t fail to catch a price drop — they let waiting inflate their budget to a price point even they no longer recognize.

7.4 Pitfalls Worth Guarding Against#

This “marginalization of gaming GPUs” is not a short-term event but a strategic pivot. Rather than waiting anxiously inside uncertain leaks, return to your real needs: buy now if you need it now, lock in the scarcest core components per your budget, and set a clear stop-loss on waiting. Don’t let one delay postpone your entire build plan indefinitely.

— Zixing, NTGame Hardware & Tech Editor