According to Uniko’s Hardware, the refreshed GPUs, including the RTX 5080 SUPER, RTX 5070 Ti SUPER and RTX 5070 SUPER, are in limbo thanks to surging memory costs. DRAMeXchange reports that DRAM and NAND flash prices have shot up, fuelled by massive demand from the AI sector, hoovering up every available wafer.

Consumer DDR5 prices have already doubled in just a few months, and GDDR spot prices are now more than twice what they were last week. While DDR7 prices are not even officially listed yet, industry insiders say they are climbing even faster.

These chips were expected to use 3 GB GDDR7 memory dies, providing up to 24 GB VRAM on the 5080 SUPER and 5070 Ti SUPER, and 18 GB VRAM on the 5070 SUPER, a tempting 50 per cent boost over their standard counterparts. But with memory prices spiralling, NVIDIA reportedly cannot produce the cards at sensible margins.

Instead, the company appears to be diverting resources to its higher-margin professional gear, such as the RTX PRO 6000 Blackwell workstation GPU and the RTX 5090 Laptop chip. Both use the same pricey 3 GB GDDR7 dies and can pack up to 96 GB VRAM in the case of the PRO 6000.

The RTX 50 SUPER range was originally tipped for an early 2026 launch, possibly debuting at CES 2026, but the latest chatter suggests that even that window is in doubt. NVIDIA has not officially confirmed any of the SUPER models, and sources close to the company hint that nothing is coming before next year at the earliest.

Nvidia can afford to wait as its GeForce RTX 50 cards face little competition from AMD’s Radeon RX 9000 lineup, and gamers are not exactly spoiled for alternatives. But with customers crying out for more VRAM and more reasonably priced GPUs, a delay or cancellation of the SUPER series could leave a big gap in NVIDIA’s consumer roadmap.