Samsung says it has developed a new GDDR6 DRAM with a 24Gbps data transfer rate, 30% faster than its 18Gbps predecessor, built on its 10nm process using EUV
The new DRAM will allow premium graphics cards to support a data processing rate of up to 1.1TB, Samsung says.
Context & Ripple Effects
Samsung’s 24Gbps GDDR6 extends its use of advanced process technology across distinct memory categories, following its 10nm LPDDR5 work while targeting the far higher bandwidth requirements of premium graphics hardware. The immediate significance is bandwidth rather than mobile-memory power efficiency.
The product also sits between GDDR6 and Samsung’s later first-generation GDDR7 development, which was positioned at 32Gbps. That makes the 24Gbps part a clear step in Samsung’s graphics-memory performance ladder rather than an endpoint.
First-order effects
- Samsung adds a 24Gbps GDDR6 option, raising the stated ceiling for premium graphics-card data processing to 1.1TB.
- Premium graphics-card designers gain a faster GDDR6 component option without changing to a new memory generation.
Second-order effects
- Samsung’s next graphics-memory generation faces a higher baseline: the subsequent 32Gbps GDDR7 roadmap is framed as an upgrade from 24Gbps GDDR6, not from the older 18Gbps tier.
- Graphics-card bandwidth competition becomes more closely tied to memory qualification and supply choices, because the reported performance gain comes from the memory component itself.
Third-order effects
- If Samsung continues moving graphics memory through successive speed tiers, premium GPU performance will increasingly depend on memory-bandwidth roadmaps as well as processor advances—a concrete instance of [[a:memory-wall|the memory wall]].
The trend: Graphics memory is advancing through shorter bandwidth steps, with GDDR6 speed improvements bridging the transition to GDDR7.