Nvidia RTX 50 SUPER GPUs delayed over 3GB GDDR7 memory chip prices
Nvidia’s RTX 50 SUPER series reportedly delayed over extortionate 3GB GDDR7 memory prices
Videocardz, citing industry sources, have stated that at least one Nvidia partner has already received RTX 50 SUPER graphics cards, implying that Nvidia’s hardware is ready. However, Nvidia has reportedly put RTX 50 SUPER production on hold due to high prices for 3GB GDDR7 chips.
Nvidia’s RTX 50 SUPER hardware is reportedly ready to manufacture. This implies that Nvidia’s RTX 50 SUPER specifications are set and are no longer subject to change. This makes sense, as early RTX 50 SUPER rumours had their launch set for Q1 2026, before the AI-induced memory shortage scuppered those plans.
According to Videocardz’s sources, 3GB GDDR7 memory chips cost $60-70, whereas a 2GB chip costs $20. In other words, the 3GB chip is 3x as expensive. That’s not worthwhile given the mere 50% increase in memory capacity it provides. In fact, Nvidia would be better off using two 2GB chips in a sandwich configuration instead of a single 3GB chip. That would lower Nvidia’s costs and increase its GPU memory capacities. However, I can’t imagine Nvidia offering gamers a 2x memory upgrade, as it could set the bar for memory capacity too high for future products. Furthermore, Nvidia doubling its use of 2GB GDDR7 chips would undoubtedly increase their prices. After all, can you name a GDDR7 memory customer aside from Nvidia?
According to our source, a 3GB GDDR7 chip currently costs between $60 and $70. A standard 2GB GDDR7 chip reportedly costs around $20. This means the 3GB version can cost about three times as much despite providing 50% more capacity.
Such pricing would add hundreds of dollars to the memory cost of a graphics card using six or eight modules.
3GB GDDR7 memory chips are what make this SUPER series SUPER
Nvidia’s RTX 50 SUPER series GPUs feature higher TDPs than their non-SUPER counterparts and use new 3GB GDDR7 memory modules. This results in higher clock speeds and a 50% increase in memory capacity for these GPUs. Only the RTX 5070 SUPER has more CUDA cores than its non-SUPER counterpart, providing greater parallelism. Simply put, the RTX 50 SUPER series doesn’t make sense without 3GB GDDR7 memory chips.
| RTX 5060 | RTX 5070 | RTX 5070 SUPER (Leaked) |
RTX 5070 Ti | RTX 5070 Ti SUPER (Leaked) |
RTX 5080 | RTX 5080 SUPER (Leaked) |
|
| SM Count | 30 | 48 | 50 | 70 | 70 | 84 | 84 |
| CUDA Cores | 3840 | 6144 | 6400 | 8960 | 8960 | 10752 | 10752 |
| Memory Capacity | 8GB | 12GB GDDR7 | 18GB GDDR7 | 16GB GDDR7 | 24GB GDDR7 | 16GB GDDR7 | 24GB GDDR7 |
| Memory Speed | 28 Gbps | 28 Gbps | 28 Gbps | 28 Gbps | 28 Gbps | 30 Gbps | 32 Gbps |
| Memory Bus Size | 128-bit | 192-bit | 192-bit | 256-bit | 256-bit | 256-bit | 256-bit |
| TGP | 145W | 250W | 275W | 300W | 350W | 360W | 415W |
Without their higher memory capacities, most of Nvidia’s RTX 50 SUPER series GPUs are simply overclocked versions of their existing counterparts. That’s not enough to merit a full product line refresh. If GDDR7 memory prices don’t decrease, these memory chips will only be viable for higher-end hardware configurations that prioritise capacity over cost-effectiveness. In other words, these chips are only viable for professional-grade products and AI-focused hardware.
Unless memory pricing decreases, the RTX 50 SUPER series may not launch anytime soon. Nvidia appears ready to produce hardware, but memory pricing remains a problem. After all, who would want an RTX 50 SUPER GPU if its pricing is crazy high?
You can join the discussion on Nvidia’s rumoured RTX 50 SUPER series delay on the OC3D Forums.
