Nvidia is reportedly exploring reduced memory configurations for its upcoming Rubin Ultra accelerator due to anticipated shortages of High Bandwidth Memory (HBM). Designs are being tested with as little as 192 GB of memory, a significant decrease from the previously announced 1 TB of HBM4E. This move may also involve a step back to HBM4 from the more advanced HBM4E, which has faced supply chain complexities. The company has stated its roadmap remains intact, but reports suggest potential delays and ongoing testing of various memory capacities and configurations. AI
IMPACT Potential HBM shortages could impact the rollout and availability of next-generation AI accelerators, affecting compute capacity.
RANK_REASON Report on potential supply chain issues impacting a major upcoming hardware release from a leading tech company.
- HBM4
- HBM4E
- Kyber NVL144
- Micron
- NVIDIA
- Nvidia GTC
- Rubin ultra
- SemiAnalysis
- SK Hynix
- The Information
- TSMC
- Tom's Hardware
AI-generated summary · Google Gemini · from 3 sources. How we write summaries →