NVIDIA is reportedly turning to Micron for memory chips used in its upcoming RTX 50 series, amid persistent challenges in the DRAM market.
NVIDIA’s RTX 50 GPUs Feature GDDR7 Modules from Multiple Suppliers, Yet Shortages Persist
The ongoing battle to procure sufficient memory chips for graphics processing units (GPUs) continues, primarily due to the relentless shortages in the DRAM sector. Historically, NVIDIA has relied on memory chips from Samsung and We Hynix for its RTX 50 series graphics cards, but new reports suggest that the company is expanding its sourcing strategy to include Micron’s GDDR7 memory modules.
Interestingly, although NVIDIA has previously sourced chips from Micron for its older GPU generations, sightings of Micron’s GDDR7 VRAM in the RTX 50 series have been extremely rare until recently. Notably, the first instance of Micron’s GDDR7 memory chips was identified in a review of the Galaxy GALAX GeForce RTX 5060 Black OC V2, detailed by Quasar Zone. The teardown revealed four Micron GDDR7 2 GB memory chips incorporated into the design.

The specific GDDR7 module identified is the MT68A512M32DF, which boasts a capacity of 16 Gb (2 GB) and operates at a speed of 28 Gbps. The notable exception in the lineup, however, is the GeForce RTX 5080, which utilizes faster 30 Gbps memory modules. The introduction of Micron’s GDDR7 components signals that NVIDIA is facing substantial difficulties in acquiring sufficient supplies from its primary memory partners.
While Micron’s involvement may help mitigate some of the supply issues for NVIDIA’s GPUs, the broader memory shortage remains a significant concern. Reports indicate that prices for NVIDIA GPUs have surged by at least 15% since October 2025, with no signs of stabilization in the near future. As such, incorporating Micron’s memory could alleviate some issues regarding availability, with Micron preparing to assist in addressing the ongoing supply crisis.
In a noteworthy development, Micron has announced its new 24 Gb GDDR7 memory modules, which are rated for an impressive 36 Gbps. However, these advanced modules will not be featured in the RTX 50 series; instead, they are slated for use in the RTX 5090 Laptop GPU. There is speculation that the upcoming NVIDIA Rubin graphics cards could also make use of these high-capacity modules.
Leave a Reply