DEVOURED
Who Eats Memory Costs?
Nvidia is passing rising memory costs onto customers, with AI server prices projected to rise by over 15% in 2027.
What: Nvidia is navigating the rising costs of HBM (High Bandwidth Memory) by adjusting accelerator pricing. While the strategy currently protects gross margins, the company expects increased content requirements to challenge its pricing power by FY28.
Why it matters: This confirms that memory supply remains the critical bottleneck for AI hardware, potentially forcing a transition in how AI infrastructure is priced and commoditized.
Deep dive
- AI server prices are expected to increase by more than 15% starting next year.
- HBM costs are becoming a significant component of final GPU and accelerator manufacturing costs.
- Nvidia's pricing strategy involves diluting the memory cost impact by bundling it within the total system value.
- FY28 is cited as a potential turning point where future memory generations may make current margin structures harder to maintain.
Decoder
- HBM (High Bandwidth Memory): A specialized, high-performance RAM interface for 3D-stacked DRAM, essential for the massive data throughput requirements of AI training GPUs.
Original article
Nvidia plans to pass rising memory costs onto customers, with AI server prices set to increase by over 15% for systems shipping next year. While HBM costs rise, Nvidia's pricing strategy helps protect gross-profit dollars by treating HBM as a smaller part of the final accelerator price. The real challenge for Nvidia will come in FY28 when new memory generations and increased content might limit its ability to maintain margin percentages.
Explore all 66 articles in today's digest
Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi