Samsung develops 8-layer HBM4E for Nvidia with 17–18Gbps target transfer speeds
AI Market Summary
Samsung's development of 8-layer HBM4E at Nvidia's request for the NVHBM platform (targeting 17–18Gbps, ~20% faster) highlights accelerating customization in AI memory and supports the roadmap for next year's Rubin Ultra GPU. The news is constructive for Nvidia's supply-chain resilience and performance-per-watt ambitions, while signaling intensifying competition in high-end HBM that could affect pricing power and qualification dynamics.
Impact level
● Medium
Affected assets
NCSKNVDA2USD/USDT-0.48%
AI Insight · NCSKNVDA2USD/USDTAI Insight
▲ Bullish
Trade now
⚠️ AI-generated insights are based on news content and are provided for informational purposes only. They do not constitute investment advice or represent the views of BingX. Investing involves risk. Please trade responsibly.
Samsung Electronics is developing an eight-layer stacked HBM4E memory chip at Nvidia’s request, aiming for transfer speeds of 17 to 18Gbps—about 20% higher than its current samples. The chip is slated for Nvidia’s custom memory platform, NVHBM, and is expected to be integrated into the Rubin Ultra AI GPU launching next year. The eight-layer design highlights a shift in AI chip development from capacity expansion toward balancing bandwidth, power consumption and yield. If the product passes validation, it could help Samsung narrow its gap with SK Hynix in high-end HBM.
