HBM, GDDR, packaging, and the interposers that move data.
5 briefs on the desk
Memory & PackagingSINGLE-SOURCE2 min brief
Leaker Claims 4 GB and 6 GB GDDR7 Memory Chips for Next-Gen GPUs
A leaker claims that 4 GB and 6 GB GDDR7 memory chips will be introduced for next-gen GPUs in the next two years.
According to Wccftech Hardware, a leaker claims that 4 GB and 6 GB GDDR7 memory chips will be rolled out for next-generation GPUs in either 2027 or 2028.…
4 GB GDDR7 chips expected6 GB GDDR7 chips expectedRollout in 2027 or 2028NVIDIA sticking with 2 GB GDDR7
The Die Brief Desk · Wccftech Hardware · September 21, 2026Read the brief →
Memory & PackagingSINGLE-SOURCE2 min brief
Samsung Plans to Increase HBM4 Chip Production by 2.5x by 2027
According to Wccftech Hardware, Samsung is set to boost its production of HBM4 memory chips by 2.5 times by 2027, driven by high demand for AI chips.
According to Wccftech Hardware, Samsung is set to boost its production of HBM4 memory chips by 2.5 times by 2027, driven by high demand for AI chips. The…
HBM4 production to increase by 2.5x by 2027Driven by AI chip demandIncreased orders for glass carriersHigher bandwidth and lower power
The Die Brief Desk · Wccftech Hardware · September 21, 2026Read the brief →
Memory & PackagingSINGLE-SOURCE1 min brief
Kioxia XL1 CXL XL-FLASH NAND Device Shown
Kioxia's XL1 is an XL-FLASH Generation 2 NAND device that attaches to a system as a 512GB memory expansion via CXL.
According to ServeTheHome, Kioxia's XL1 is an XL-FLASH Generation 2 NAND device that attaches to a system as a 512GB memory expansion device via CXL. The mechanism is…
The Die Brief Desk · ServeTheHome · September 20, 2026Read the brief →
Memory & PackagingSINGLE-SOURCE2 min brief
University of Michigan Unveils Chiplet Framework for AI Accelerators
University of Michigan researchers published 'Fengshui,' a chiplet co-design framework that jointly optimizes chiplet pool composition and bespoke ASIC design for AI accelerators.
According to Semiconductor Engineering, University of Michigan researchers published a technical paper titled 'Fengshui: Demystifying Chiplet Ecosystem and Bespoke Neural Network Accelerator Codesign.' The framework jointly optimizes chiplet…
Fengshui frameworkChiplet pool co-designBespoke ASIC designEnergy + cost reduction
The Die Brief Desk · SemiEngineering · September 20, 2026Read the brief →
Memory & PackagingSINGLE-SOURCE2 min brief
Heidelberg University Advances Chiplet-Based Neuromorphic Computing
Heidelberg University researchers published a paper on a unified interconnection network for chiplet-based scaling of the BrainScaleS neuromorphic system.
According to Semiconductor Engineering, Heidelberg University researchers published a technical paper titled 'A Unified Interconnection Network for Chiplet-Based Scaling of the BrainScaleS Neuromorphic System.' The paper addresses the…