Key Takeaways
- The AI-driven memory expansion remains in its infancy and may surpass the 2001-2007 NAND growth period, according to Citi analysts
- Leading memory manufacturers Micron, Samsung, and SK Hynix have declined over 20% from peak levels
- Citi strategists are accumulating positions in AMD, Texas Instruments, and Applied Materials during the correction
- HBM supply constraints are anticipated to drive AI infrastructure toward higher GPU deployment densities
- Major cloud infrastructure providers’ capital expenditure projected to surge 90% in 2026, bolstering semiconductor demand
The memory chip sector has experienced significant turbulence in recent weeks. Leading players including Micron, Samsung, and SK Hynix have retreated more than 20% from their recent highs.
The correction followed an impressive rally, with investor anxiety centered on elevated valuations and questions surrounding artificial intelligence infrastructure investment sustainability. However, Citi’s research team characterizes this downturn as a strategic accumulation point rather than a red flag.
According to Citi’s analysis, the artificial intelligence-fueled memory expansion cycle remains in its nascent phase. Their framework draws parallels to the 2001-2007 NAND expansion, when emerging consumer electronics such as MP3 players and digital cameras sparked unprecedented demand.
The current environment differs significantly, with AI applications simultaneously driving requirements for both DRAM and NAND technologies. Citi’s research suggests this dual-demand dynamic positions the present cycle to exceed the performance of its predecessor.
Further evidence of cycle durability comes from enterprise customers entering multi-year supply agreements spanning three to five years. Such long-term commitments signal confidence in persistent demand rather than temporary momentum.
HBM Supply Constraints Driving AI Architecture Evolution
The ongoing scarcity of high-bandwidth memory components is another critical factor. Citi anticipates this constraint will drive AI chip manufacturers toward architectural strategies that emphasize greater GPU quantities with reduced memory allocation per processing unit.
Despite lower per-chip memory configurations, Citi forecasts aggregate HBM capacity within AI infrastructure will expand by 434%, climbing from approximately 20 terabytes to beyond 110 terabytes, as system-level GPU counts increase from 72 to 576 units.
SK Hynix disclosed during its second-quarter financial update that management is evaluating shareholder return initiatives. Citi anticipates an announcement prior to third-quarter results and maintains its Buy recommendation with a 3,100,000 won price objective.
Citi Identifies Additional Semiconductor Opportunities
Beyond memory manufacturers, Citi is establishing positions in three semiconductor stocks amid the pullback: AMD, Texas Instruments, and Applied Materials.
The Philadelphia Semiconductor Index has climbed nearly 60% year-to-date but retreated 21% during the current quarter. Citi attributes the decline to elevated investor expectations colliding with market realities.
Data center semiconductor consumption remains robust, representing 34% of total market activity. Industrial applications are expanding 30 to 35% year-over-year, while automotive semiconductor demand is advancing 12 to 15%.
Certain analog chip categories now show delivery lead times exceeding 16 weeks. Customer urgency indicators have doubled, signaling constrained supply conditions.
Citi elevated its capital expenditure growth projections for the five largest cloud service providers to 90% and 46% for 2026 and 2027 respectively. Alphabet increased its 2026 capex guidance to a range of $195 billion to $205 billion. Amazon raised its projection to $220 billion.
Citi views this environment as fundamentally supportive for semiconductor equities and continues recommending AMD as its preferred position, emphasizing the company’s GPU and CPU market share expansion. Applied Materials receives positive catalyst coverage ahead of its August 13 earnings disclosure.
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
