Nvidia And SK Group Tease A Deal As Memory Stays Tight - Finimize

The potential collaboration between NVIDIA and SK Group signals a significant development in the semiconductor industry amid persistent memory chip shortages. This partnership reflects the industry's ... Read original →

The rumored NVIDIA-SK Group alliance is a structural response to surging AI compute demand colliding with persistent HBM shortages. Technically, it will accelerate co-optimization of CoWoS packaging and HBM3E stacking, pressuring TSMC and Samsung to advance interposer and TSV processes while eroding Micron’s pricing leverage in AI memory. On compliance, U.S. export controls are forcing both parties to build supply chain redundancy—SK Hynix’s Wuxi fab acts as a critical buffer but risks secondary sanctions. Competitively, AMD will likely deepen vertical integration with Japanese/Korean material suppliers, while Intel may leverage its IFS platform to onboard third-party memory makers for Gaudi ecosystem viability. Over the next 18 months, such 'compute-memory' alliances will evolve from tactical fixes to strategic norms, shifting the industry from fabless-foundry separation toward system-level co-definition—and ultimately redrawing the power map of the advanced semiconductor supply chain.