Please note that this post is tagged as a rumor .
NVIDIA GR200 GPU series for next-gen GeForce
GR212 is not for gamers.
We are midway through CES 2026, which is typically NVIDIA’s place to announce new cards. This time, NVIDIA was expected to launch the RTX 50 SUPER series, a mid-cycle refresh for the Blackwell family. However, it looks like the memory shortage, and possibly other factors, have persuaded NVIDIA to wait.
The leaker who posted the first RTX 50 SUPER rumors last year has been quiet for a while. Still, it seems Kopite7kimi has an update, and it points to the next generation. What is almost certainly the RTX 60 series is now said to use the Rubin GPU architecture, the same one NVIDIA is currently highlighting in its data center presentations.
GR20x is for gaming.
— kopite7kimi (@kopite7kimi) January 7, 2026
As we recall, NVIDIA already announced a large Rubin CPX GPU that was apparently based on Rubin and featured CUDA cores. Naturally, rumors started swirling that this could be the next-gen gaming GPU as well. Today, Kopite7kimi put those rumors to rest and said the GPU for that product is GR212, which is not the gaming variant.
Apparently, the GR200 series will be what NVIDIA uses for the RTX 60 series. This suggests Rubin is coming to the GeForce lineup, which had not been clearly confirmed until now.
It's NOT a giant for gaming. It is GR212. As you see, GR100 is based on two GR102, looks the same as GB100. https://t.co/4Pgdjv6Wt6
— kopite7kimi (@kopite7kimi) January 7, 2026
Since we are talking about something with no real leak trail behind it, we might as well speculate. Rubin should use a more advanced node. That gives NVIDIA two options: keep die size roughly where it is, or reduce it. Going larger does not make much sense for yields. Also, GB202 was the only part that saw the biggest jump in both core count and die size. Transistor density is similar between Ada and Blackwell. A more advanced node could push density higher, though.
GeForce flagship GPU sizes:
- GR202 Rubin ⬅️
- GB202 Blackwell : 750/761 mm²
- AD102 Ada : 608 mm²
- GA102 Ampere : 628 mm²
- TU102 Turing : 754 mm²
- GV100 Volta : 815 mm²
- GM200 Maxwell : 601 mm²
- GK110 Kepler : 551 mm²
A typical expectation is around a 30% jump per generation, and that often holds for flagship SKUs. Rubin-based GeForce GPUs are unlikely to arrive before 2027. If the RTX 50 SUPER series slips into Q2 (not confirmed), that also suggests RTX 60 will not land any sooner than about a year later.
| NVIDIA RTX 40/50/60 GPUs | ||||
|---|---|---|---|---|
| VideoCardz | SM Count | Die Size | Transistors | SKU |
| RUBIN | ||||
| GR202 | TBC | TBC | TBC | RTX 6090-class |
| GR203 | TBC | TBC | TBC | RTX 6080-class |
| GR205 | TBC | TBC | TBC | RTX 6070-class |
| GR206 | TBC | TBC | TBC | RTX 6060-class |
| GR207 | TBC | TBC | TBC | RTX 6050-class |
| BLACKWELL | ||||
| GB202 | 192 | 750mm² | 92.2B | RTX 5090(D) |
| GB203 | 84 | 378mm² | 45.6B | RTX 5080/5070Ti |
| GB205 | 50 | 263mm² | 31.0B | RTX 5070 |
| GB206 | 36 | 155mm² | 16.9B | RTX 5060 Ti/5060 |
| GB207 | 20 | 181mm² | 21.9B | RTX 5050 |
| ADA LOVELACE | ||||
| AD102 | 144 | 608mm² | 76.3B | RTX 4090(D) |
| AD103 | 80 | 379mm² | 45.9B | RTX 4080(S)/4070TiS |
| AD104 | 60 | 295mm² | 35.8B | RTX 4070(S)/Ti |
| AD106 | 36 | 188mm² | 22.9B | RTX 4060Ti |
| AD107 | 24 | 159mm² | 18.9B | RTX 4060 |
Source: @kopite7kimi
