Regarding NVIDIA's unannounced GPU 'GR20x,' prominent leaker Kopite7kimi has suggested it will be a 2028 product. This pushes back the previous estimate of late 2027, but it is not a delay to any schedule NVIDIA has officially announced. The company has not revealed RTX 60, GR20x, or a consumer version of Rubin, and in its response to Tom's Hardware, it stated that it does not comment on rumors or unannounced products. Still, the year 2028 carries significant weight. If it holds true, the RTX 50 generation would remain the current generation for at least seven months longer than any of the previous three generations.
The 2028 Theory for GR20x and the Boundary With Official Roadmaps
On September 21, 2026, Kopite7kimi posted that GR20x is a '2028 product,' indicating a belief that it had been delayed again. The post also stated that updates would become less frequent until reliable news emerges. Since the original X post could not be directly accessed, this article's published content is limited to what Tom's Hardware verified from the post and its own independent inquiry to NVIDIA.
What's worth noting here is that the baseline for 'delay' is itself based on the leaker's own prediction. NVIDIA has never officially announced the earlier late-2027 timeline. Therefore, the only confirmed change is that the same source revised its own forecast from late 2027 to 2028. Whether the product plan actually shifted internally cannot be verified.
The name 'Rubin' also does not, by itself, corroborate anything about GeForce. NVIDIA's GTC 2025 materials describe Rubin as a next-generation platform designed to power trillion-parameter models and large-scale AI factories. In the latest 10-Q, Vera Rubin also appears as the next-generation data center architecture, but no correspondence with GeForce, GR20x, or RTX 60 is indicated. Even when the same year '2028' or the name 'Rubin' appears in different contexts, readers need to distinguish between NVIDIA's official AI infrastructure roadmap and unverified consumer product information.
If 2028 Holds, the RTX 50 Generation Would Last at Least 35 Months
Measured by NVIDIA's official launch dates, the top-tier desktop GeForce RTX products have transitioned between generations roughly every 24 to 28 months over the past three cycles. If the leak that GR20x is a 2028 product proves accurate, the gap from RTX 50's January 30, 2025 launch would be at least approximately 35 months.
For comparison, the benchmark used is the launch date of the first top-tier desktop product released in each generation. Announcement dates and launch dates were not mixed, and since the exact month and day for 2028 are unknown, January 1 was used as the earliest possible placeholder.
| Generation | Reference Product | Launch Date | Gap From Previous Generation |
|---|---|---|---|
| RTX 20 | RTX 2080 Ti / 2080 | September 20, 2018 | — |
| RTX 30 | RTX 3080 | September 17, 2020 | 728 days (23.9 months) |
| RTX 40 | RTX 4090 | October 12, 2022 | 755 days (24.8 months) |
| RTX 50 | RTX 5090 / 5080 | January 30, 2025 | 841 days (27.6 months) |
| GR20x 2028 theory | NVIDIA unannounced | On or after January 1, 2028 | 1,066+ days (35.0+ months) |
The number of months was calculated by dividing the number of days by 365.25÷12. If the actual timing within 2028 turns out to be later than January 1, the gap would be even longer. That said, the product tier of the first-released item in each generation is not perfectly consistent, and this calculation does not prove the existence of GR20x or confirm its product name. What it does show is the magnitude by which the 2028 theory deviates from past launch intervals.
The longest interval among the most recent three generations was the 27.6 months from RTX 40 to RTX 50. The 35.0-month figure exceeds that by at least 7.4 months. For buyers planning to wait for the next generation, this means the usual roughly two-year cycle may no longer apply as expected.
The Scale Gap in AI Business Is Context, Not Proof of a Delay
In NVIDIA's fiscal 2027 second quarter, Data Center revenue was $89.023 billion, while Edge Computing revenue was $7.198 billion. A simple division puts the ratio at 12.37 times. However, the Edge Computing category, introduced starting in fiscal 2027's first quarter, is not a gaming-exclusive segment. It also includes Blackwell workstations, and prior-year comparison figures have been restated to match the new classification.
The same 10-Q states that consumer PC sales slowed due to rising memory and system prices. Additionally, supply and manufacturing capacity commitments grew from $119.0 billion in the previous quarter to $279.0 billion as of July 26, 2026. These commitments primarily cover memory and manufacturing equipment for data center infrastructure, and NVIDIA also explains that it faces certain supply constraints while shipping Blackwell and Rubin in parallel.
It can be confirmed that NVIDIA's AI business is vastly larger and that the company holds enormous supply commitments. However, these figures alone cannot support the conclusion that 'GR20x was delayed because AI was prioritized.' The 10-Q does not disclose individual consumer GPU plans or how design resources were allocated. Regarding GDDR7 pricing, advanced manufacturing processes, and plans for the RTX 50 SUPER, the primary sources reviewed for this article likewise offer no indication of delay reasons.
For Purchase-Delay Decisions, Watch Official Milestones, Not the Year
The 2028 theory alone provides weak grounds for postponing a needed GPU purchase for an extended period. NVIDIA has not confirmed not only the launch timing but even the product name 'RTX 60' or its correspondence with GR20x. Even if the leak is accurate, it remains unknown which month of 2028 it would land in, whether an interim update would be inserted, or what the pricing and memory capacity would look like.
There are three milestones for a purchase decision. First, NVIDIA must announce the product name and architecture. Second, it must indicate mass production or a launch date. Third, buyers need to be able to compare the new product's price and memory configuration against the street price of RTX 50 at that time. At this stage, only two things are certain: that the top-tier desktop RTX 50 product launched on January 30, 2025, and that if the 2028 theory proves true, the current generation's lifespan would clearly exceed the previous three generations.
Whether to wait for the next generation or choose a current product cannot be decided by the single word '2028.' Only once NVIDIA officially links a consumer product's name to a launch date can wait time be placed on the same table alongside performance and price.