Gelsinger calls HBM 'lousy', SK Hynix says it's not memory's final answer

Industry leaders are questioning the long-term viability of High Bandwidth Memory (HBM) for AI. Despite its current importance, Intel's CEO and SK Hynix suggest alternative memory solutions are needed for future AI hardware advancements.
Key takeaways
- HBM, crucial for AI chips, faces criticism from industry figures.
- Intel CEO and SK Hynix signal HBM is not the ultimate solution.
- New memory technologies may be essential for future AI hardware.
- This could influence AI hardware development and accessibility.
Why it matters
This discussion impacts AI tool developers and users by highlighting potential shifts in hardware requirements. Future AI performance and cost could depend on the adoption of new memory technologies beyond current HBM standards.
Try this on SynaBot
Related AI assistants, prompts, and tools from the SynaBot catalog.
- FinalScoutFinalScout works in the AI space. Its core capability is AI-driven email finding and crafting with LinkedIn integration, so it fits teams that want to support a variety of AI-assisted workflows across business and personal use cases.
- FinalleFinalle — Financial Intelligence Reimagined. It sits in the AI category and is built to support a variety of AI-assisted workflows across business and personal use cases.
- Final TouchFinal Touch revolutionizes product visuals for businesses by offering AI-driven scene generation and no-code editing, enabling teams to create stunning marketing assets efficiently.


