February 2026 ISSCC Conference Teases High-Speed Innovations: GDDR7, LPDDR6, HBM4 & More

The IEEE ISSCC 2026 conference is poised to take center stage this February in San Francisco, showcasing a plethora of innovations in technology. Spanning five days, this prestigious event will unveil cutting-edge research and developments in the realm of solid-state circuits and beyond, drawing attention from industry leaders and enthusiasts alike. Among the plethora of topics, the spotlight will be on groundbreaking advancements in DRAM technology, including GDDR7, HBM4, and LPDDR6.

Spotlight on DRAM Innovations

IEEE’s upcoming Solid-State Circuits Conference promises a comprehensive lineup of sessions, with significant focus on DRAM advancements. Noteworthy contributions come from SK hynix, which is set to reveal its new GDDR7 DRAM with impressive 48 Gbps speeds, achieving a remarkable 70% improvement over current models. SK hynix’s innovations promise up to 192 GB/s bandwidth, a leap forward for tech enthusiasts eagerly awaiting the next-gen possibilities this development heralds. Alongside, Samsung’s HBM4 and LPDDR6 papers will further delve into future memory solutions, with Samsung’s HBM4 offering 3.3 TB/s bandwidth per stack, tailored for NVIDIA’s future AI accelerators.

A schedule lists presentations on DRAM advancements from Samsung Electronics and SK hynix, including A 36GB 3.7TB/s HBM4 DRAM with Per-Channel TSV RDQS at 10:05 AM and A 48Gb/s 24Gb GDDR7 DRAM for Mid-Range Inference AI
Image Source: ISSCC

In the realm of LPDDR6, SK hynix will present findings on its 1cnm LPDDR6 SDRAM, achieving up to 14.4 Gbps pin speeds. Samsung is not far behind, with its LPDDR6 research showing 12.8 Gbps rated speeds. Additionally, Samsung’s HBM4 DRAM flaunts 36 GB capacity with unprecedented bandwidth, marking its alignment with NVIDIA’s next-gen AI technologies.

GDDR Graphics Memory Evolution

GRAPHICS MEMORY GDDR7 GDDR6X GDDR6 GDDR5X
Workload Gaming / AI Gaming / AI Gaming / AI Gaming
Platform (Example) GeForce RTX 5090 GeForce RTX 4090 GeForce RTX 2080 Ti GeForce GTX 1080 Ti
Die Capacity (Gb) 16-64 8-32 8-32 8-16
Number of Placements 12 12 12 12
Gb/s/pin 28-48 19-24 14-16 11.4
GB/s/placement 112-192 76-96 56-64 45
GB/s/system 1536-2304 912-1152 672-768 547
Configuration (Example) 384 IO (12pcs x 32 IO package)? 384 IO (12pcs x 32 IO package) 384 IO (12pcs x 32 IO package) 384 IO (12pcs x 32 IO package)
Frame Buffer of Typical System 24 GB (16 Gb)
36 GB (24 Gb)		 24 GB 12GB 12GB
Module Package 266 (BGA) 180 (BGA) 180 (BGA) 190 (BGA)
Average Device Power (pJ/bit) TBD 7.25 7.5 8.0
Typical IO Channel PCB (P2P SM) PCB (P2P SM) PCB (P2P SM) PCB (P2P SM)

These sessions promise deep dives into highly specialized topics, with giants like Intel, NVIDIA, and AMD ready to unveil their own cutting-edge research. The ISSCC 2026 conference is set to be a landmark event for those keen on witnessing the forefront of technological evolution in DRAM and beyond.