The race for faster and more efficient graphics memory is heating up with the upcoming GDDR7. This next-gen memory, capable of reaching speeds of up to 40 gigatransfers per second—about twice as fast as today’s GDDR6X—will commence mass production in Q3 2024, as announced by SK Hynix.
While previous projections placed GDDR7’s arrival in early 2025, aligning with Nvidia’s anticipated RTX 5000 series release, the early start in production could mean that new GPUs might launch sooner. The availability of GDDR7 could potentially alleviate some of the common supply issues seen with new GPU releases, particularly given the growing demand from AI sectors.
Both Micron and Samsung, alongside SK Hynix, have demonstrated GDDR7 samples this year. Despite the fastest chips being rated at 32Gbps with 16GB modules, this still represents a significant leap over the current 21Gbps GDDR6X memory used in the RTX 4090.
While we await further details on the specifications of upcoming GPUs from Nvidia, AMD, and Intel, the arrival of GDDR7 points to a future where graphics performance will reach new heights. Given the high demand for powerful GPUs, especially in AI applications, these next-gen chips are likely to come at a premium.