ChipsThe story, in brief

Maybe the chips are down because memory’s too expensive

DRAM costs just became the silent bottleneck killing AI infrastructure ROI.

Paper-cut illustration of an amber microchip with circuit paths extending into a row of data-center cabinets.
The infrastructure powering AI.AI illustration by KeyNews
The KeyNews take

Why it matters

Memory pricing is constraining AI chip deployment economics. As GPU costs plateau, DRAM expenses are emerging as the underestimated cost driver limiting large-scale model inference and training infrastructure buildouts.

The key facts

4 to know
  1. DRAM pricing pressure identified as infrastructure constraint

  2. Memory costs competing with GPU economics for AI capex budgets

  3. July 2026 analysis of chip supply chain cost structure

  4. Inference/training deployment economics impacted by memory component pricing

Go to the source

Financial Times Technologyft.com

Publisher excerpt: DRAM, DRAM, DRAM / Did I ruin your plans, plans, plans?
Read original report
Back to today's editionMore chips news

Keep reading

Related stories

More from Chips