NVIDIA RTX 5090 Laptop GPU Expected to Feature 24 GB GDDR7 Memory, Signifying Major Generational Upgrade

NVIDIA is set to enhance its GeForce RTX lineup with the highly anticipated RTX 5090, expected to feature a substantial increase in VRAM compared to its predecessor and significantly improved memory bandwidth.

Enhanced VRAM and Memory Type for the GeForce RTX 5090

Recent leaks have shed light on the GeForce RTX 5090’s specifications for the mobile platform. The new GPU will maintain the same VRAM capacity as the RTX 4090, offering 16 GB of memory. However, the standout feature of this upgrade is its transition to GDDR7 memory, an advancement over the GDDR6 found in previous models.

This switch to GDDR7 memory is projected to remarkably increase the memory bandwidth, setting a new standard for performance. Previously, the RTX 4090 could not be equipped with more VRAM due to the lack of suitable 3 GB GDDR6 modules. The upcoming RTX 5090 is rumored to be the first GPU to utilize 3 GB GDDR7 memory modules, as reported by the well-known leaker Golden Pig Upgrade (via Weibo).

Golden Pig Upgrade RTX 5090 Laptop 3GB GDDR7

Advancing to enhance memory capacity, the RTX 5090 could escalate its VRAM to 24 GB while retaining the existing 256-bit memory bus architecture. This adjustment would provide an 8 GB boost over the confirmed 16 GB version of the RTX 5090 laptop edition, aligning it more closely with the desktop RTX 4090 model in terms of memory capacity. While such an increase may not be necessary for all gaming scenarios, it could significantly benefit gamers playing modern, memory-intensive titles at 4K resolution, especially those utilizing mods.

Leading Performance in the Blackwell Series

Positioned as the flagship GPU of NVIDIA’s Blackwell series, the GeForce RTX 5090 laptop variant is poised to be the fastest within this product line. As things currently stand, no other GPU is slated to incorporate 3 GB GDDR7 memory modules, elevating the RTX 5090 laptop GPU to a distinct position in the market.

While NVIDIA has yet to release official specifications, leaks indicate that the RTX 5090 will utilize remarkable 28 Gbps GDDR7 memory, a significant enhancement over the 18 Gbps GDDR6 found in the RTX 4090. The mobile variant of the RTX 5090 is also expected to employ the GB203 die, differing from NVIDIA’s typical GB202 choice for mobile GPUs.

On the performance front, the desktop variant of the RTX 5090 is anticipated to feature a striking 21,760 CUDA cores, while the mobile version will likely include approximately 10,752 cores, being roughly half that of the desktop model. With a maximum thermal design power (TDP) of 175W, it is expected to demand slightly more power than its predecessor. Recent sightings of the RTX 5090 laptop GPU have been reported in forthcoming ASUS ROG laptops, among other models from the RTX 50 series.

Source&Images