The flagship spot for gaming graphics cards now has to be filled by a shelved old design.

Reports state that the GB202 die has been prioritized for higher-margin professional GPUs, leaving the gaming lineup without a flagship; the once-shelved RTX 5080 24GB plan has been revived as a forced new flagship, with an expected launch window in Q1 2027.

In terms of specs, it retains the GB203 GPU die with 10,752 CUDA cores; VRAM is upgraded from 16GB GDDR7 to 24GB while maintaining a 256-bit bus width, with a total board power of around 415W.

Notably, this is not a brand-new design. It was originally planned as the RTX 5080 SUPER, but the project was put on hold due to skyrocketing memory costs. It has now been brought back for a clear purpose: to fill the vacancy in high-end gaming GPUs.

As flagship models such as the RTX 5090 exited the market, a gap opened at the top of NVIDIA’s gaming GPU stack. Combined with soaring memory costs and AI demand squeezing manufacturing capacity, the supply and pricing logic for high-end gaming graphics cards has been completely rewritten.

A simple calculation makes this clear. The same GB202 die inside an RTX 5090 gaming card can fetch more than triple the price when fitted into an RTX Pro 6000. Advanced process capacity has long been tight. It is easy to guess whether limited wafers will go to enterprise products priced at tens of thousands of US dollars or consumer cards in the $2,000 bracket.

Current leaks indicate: The 16GB variant launches at $999; retail pricing for the 24GB version is expected to range from $1,999 to $2,999, roughly equivalent to 14,300 to 21,500 CNY.

The domestic channel picture is more straightforward: The actual selling price of the 16GB model has climbed from the 8,299 CNY reference MSRP to roughly 12,000 CNY, and the 24GB variant will likely exceed 15,000 CNY directly.