Memory has grown to nearly two-thirds of AI chip component costs
Summary
Memory costs have risen to 63% of AI chip component costs between 2024 and 2025, driven by HBM spending, according to analysis from Epoch AI. Total component spending on AI chips grew from $22 billion to $52 billion over the same period.
View Cached Full Text
Cached at: 05/24/26, 06:40 PM
Similar Articles
Memory is now 63% of what an AI accelerator costs to build, up from 52% in early 2024
Memory now accounts for 63% of AI accelerator costs, up from 52% in early 2024, shifting industry priorities toward data movement optimization over process shrinks.
Who Eats Memory Costs? (9 minute read)
The article analyzes how rising HBM memory costs are affecting NVIDIA's AI server pricing and gross margins, suggesting that customers may bear the increased costs for upcoming systems like Vera Rubin and Grace Blackwell.
On-device AI hardware is no longer a viable business with excessive RAM prices
A developer of on-device AI hardware says rising memory and component prices (16GB LPDDR5X + 128GB NAND alone exceeding $300) may force device pricing above $1,500, undermining the business model despite positive customer feedback, and asks other AI hardware builders for strategies.
@rohanpaul_ai: The AI capex cycle is now pushing memory into a much more aggressive pricing upcycle than expected. UBS now expects DRA…
AI infrastructure investment is driving up demand for memory (DRAM, NAND), causing aggressive price increases. UBS forecasts DRAM and NAND price hikes in Q3 and Q4, with the DRAM market expected to remain undersupplied until at least 2028.
The AI boom is showing up on price tags
AI data centers are consuming memory chips, driving up prices for phones, laptops, and consoles. Major manufacturers like Apple, Sony, Microsoft, and Nintendo have raised prices, and further increases are expected.