Memory supply will remain tight through 2028

Memory is not a new AI infrastructure bottleneck.

The warning signs have been visible throughout 2026.

In January, Reuters reported sharp memory inflation and planned PC price increases of up to 20% from major vendors.

IDC linked the shortage to AI data centers consuming more memory capacity, placing pressure on DRAM and NAND supplies.

Micron Technology reinforced the trend throughout the year:

• March: Record results highlighted memory’s strategic value in the AI era
• June: Long-term customer agreements reflected demand exceeding supply
• September: Customer commitments reached $32 billion, with agreements covering most of Micron’s 2027 HBM supply

Micron now expects tighter memory and storage conditions in fiscal 2027 and 2028 than in 2026.

New production capacity will take time to provide meaningful relief.

The September 30 announcement does not introduce a new issue.

It confirms a longer and broader constraint.

The impact extends beyond AI clusters:

• Higher server and storage costs
• Longer delivery times
• Greater pricing volatility
• Pressure on PC and device prices
• Increased deployment risk

My takeawayMicron’s results reinforce a constraint visible throughout 2026.

The question is no longer whether memory will limit AI infrastructure growth.

The question is how long the constraint will last and how broadly higher costs will affect technology spending.

Is your infrastructure plan accounting for memory availability and pricing through 2028?