Nvidia raises AI server prices 15% as memory becomes new bottleneck
heyshrutimishra · x · 2026-08-24
Nvidia has notified Microsoft, Google, and Oracle that server prices based on Grace Blackwell and Vera Rubin chips will rise by over 15% starting early next year.
Cause: The increase is driven by soaring costs for High-Bandwidth Memory (HBM) and server DRAM, not chip manufacturing costs. Nvidia is passing these costs to clients despite maintaining 75% gross margins.
Implications: The AI infrastructure bottleneck is shifting from GPUs to memory. Control of the memory supply chain now lies with SK Hynix, Samsung, and Micron, effectively making them the new toll booth for AI buildout. Nvidia reports earnings this Wednesday.
Related event: Nvidia to Raise AI Server Prices Over 15% as Memory Costs Surge(12 posts)→
More from Infra
- US export controls may force China to eliminate Nvidia dependency — VraserX · 2026-08-24
- xllm generates an image in 0.4 seconds — warycat · 2026-08-24
- WULF CEO reveals modern AI data centers use minimal water via closed-loop systems — robleclerc · 2026-08-24
- Cursor Team Publishes 'Git at Any Scale', Advocating for Stateless Infrastructure — thesephist · 2026-08-24
- AI Performance Engineering resource list v2 covers everything from CUDA to MoE serving — AccBalanced · 2026-08-24
- Semiconductor engineers now more prestigious than doctors in South Korea — SuB8u · 2026-08-24