Next-Gen NVIDIA & AMD GPUs to Utilize GDDR7 Memory

Insider information suggests that future NVIDIA and AMD graphics cards may feature 16 GB dies or larger with GDDR7 memory.
Next-Gen NVIDIA & AMD GPUs to Utilize GDDR7 Memory

Highlights

  • NVIDIA and AMD poised to launch GPUs with GDDR7 memory, delivering enhanced performance in demanding PC games.
  • While AMD is advancing with Radeon GPUs, NVIDIA maintains market dominance with GeForce RTX cards.
  • Both next-gen NVIDIA and AMD GPUs stand to benefit from increased memory bandwidth and larger memory per die.

According to a trusted source in the tech industry, the next wave of NVIDIA and AMD GPUs is expected to incorporate GDDR7 memory with 16 GB dies or higher. This development could lead to a significant performance boost for upcoming PC games that demand high graphical capabilities.

Although AMD has made notable progress with its Radeon RX 7800 XT and 7900 GRE cards, NVIDIA remains the leading player in the PC gaming sector. NVIDIA's GeForce RTX series is popular among both gamers and professionals due to its superior ray tracing capabilities and support for cutting-edge technologies like DLSS 3.0. Meanwhile, AMD continues to innovate with technologies such as FidelityFX Super Resolution 3.0 and HYPR-RX, positioning both NVIDIA and AMD to leverage the benefits of increased memory bandwidth and larger memory per die in their next-generation GPUs.

Recent reports from Wccftech reveal that the initial lineup of GDDR7 NVIDIA and AMD GPUs will feature 16 GB dies, each with a minimum of 2 GB of VRAM per module. This marks a significant upgrade from GDDR6 GPUs, which debuted with 8 GB dies and 1 GB of VRAM per module. The transition to GDDR7 suggests that the new NVIDIA and AMD GPUs will likely offer higher VRAM capacities across their entire range compared to predecessors like NVIDIA's RTX 4060 Ti and AMD's RX 7600. However, it's important to note that the memory configurations of the first-generation GDDR7 GPUs may closely resemble those of current-gen GPUs as manufacturing processes evolve over time.

NVIDIA and AMD First Gen GDDR7 GPU Bandwidth and Memory Projections

  • 128-bit @ 32 Gbps: 512 GB/s (8 GB)
  • 192-bit @ 32 Gbps: 768 GB/s (12 GB)
  • 256-bit @ 32 Gbps: 1024 GB/s (16 GB)
  • 320-bit @ 32 Gbps: 1280 GB/s (20 GB)
  • 384-bit @ 32 Gbps: 1536 GB/s (24 GB)
  • 512-bit @ 32 Gbps: 2048 GB/s (32 GB)

Major chip manufacturers like Samsung and Micron are driving the GDDR7 technology forward. While NVIDIA and AMD are expected to introduce GPUs with 24 GB or larger dies, it might require another generation or a significant refresh, such as NVIDIA's RTX 40 Super series, for higher memory configurations to become widely available. GPUs with 24 GB dies could potentially support up to 3 GB of VRAM per module, while 32 GB dies may accommodate 4 GB per module. These advancements could materialize in new GPU releases as early as 2026.

The future of GPU technology and PC gaming looks promising, with the prospect of high-performance GPUs offering ample video memory at competitive price points. Both NVIDIA and AMD are anticipated to implement substantial architectural enhancements while transitioning to GDDR7 memory. However, optimal gaming experiences will also require high-performance DDR5 RAM to complement the new GPUs.

  • Sapphire AMD Radeon NITRO+ RX 7900 XTX Vapor-X

    The Sapphire Nitro lineup is synonymous with the highest-end AMD 3rd-party GPU offerings in the market. The Nitro+ doubles down on this with the highest-in-market clocks and a vapor chamber that outperforms the stock design considerably while trading at a significant premium over competing 7900XTX-based products.

  • ASUS TUF Gaming NVIDIA GeForce RTX™ 4070 Ti OC Edition
    $749 $860 Save $111

    The Asus TUF RTX 4070TI is a silent but top-tier performer that builds on the TUF lineup's advantages while trading at a minor premium from Nvidia's stock Founders Edition that brings no-nonsense overclocking to a very capable vapor chamber to maintain tight thermal control over excellent performance.