AI Data Centres Are Creating a Memory and Storage Bottleneck – Here’s What Indian Investors Should Know

When artificial intelligence (AI) became a dominant theme in technology and capital markets, most headlines centered on processors and cloud platforms. Yet a more foundational story was unfolding underneath, i.e., the sharp rise in memory and storage demand driven by AI data centres. This trend has created an unusual supply crunch in memory markets, pushing some U.S.-listed companies into rare stock market leadership positions in 2025 and early 2026.

For Indian investors using platforms such as Appreciate to access U.S. equities, this structural shift represents an under-appreciated segment of the AI infrastructure stack and a potential long-term investment theme.

Memory: A Highly Concentrated Global Market

Unlike other areas of semiconductors, memory markets are extremely concentrated. In 2025, essentially three major players dominated the global DRAM landscape.

SK hynix, Samsung Electronics, and Micron account for over 90% of global DRAM output, with SK hynix emerging as the largest manufacturer in recent quarters.

In Q2 2025, TrendForce estimated SK hynix at 38.7%, Samsung at 32.7%, and Micron at 22% of global DRAM revenue.

Global DRAM Market Share (Q2 2025)

Memory isn’t a commodity with dozens of players. These three companies set much of the global pricing  and capacity allocation, meaning that any structural demand shift impacts prices quickly and broadly.

For NAND flash (the non-volatile storage used in SSDs), the field is slightly broader but still dominated by a handful of firms: Samsung, SK hynix, Micron, YMTC, and Western Digital.

Global NAND Flash Market Share (2025)

AI Data Centres: A Structural Demand Driver

AI workloads are far more memory-intensive than typical compute tasks. Large language models and other generative AI systems require huge volumes of DRAM and flash to train and serve, driving disproportionate memory use.

Cloud providers and hyperscalers such as Amazon, Microsoft, Google, and Meta have significantly expanded AI-related data-centre spending, contributing to a record build-out of U.S. server infrastructure in 2025. An S&P Global report noted that U.S. data-centre construction spending reached $61 billion in 2025, a level tied directly to AI demand.

Many memory manufacturers are pivoting production capacity toward high-bandwidth memory (HBM), a key component of AI servers, at the expense of traditional consumer and commodity memory. This reduces the supply of DRAM and NAND available for other segments.

The result is a structural mismatch: rapidly growing enterprise consumption versus slow-to-expand global capacity.

A Memory Supply Crunch Is Unfolding

Global DRAM and NAND markets experienced cyclical oversupply during the pandemic and early 2020s. But beginning in 2024 and continuing through 2025 and 2026, a new pattern emerged.

Memory pricing, particularly for server and enterprise memory, has risen sharply. Industry trackers report DRAM contract price increases of 15% or more month-on-month into late 2025, with NAND rising similarly.

Some analyst projections suggest that DRAM prices could double by 2026 compared with early 2025 levels, driven by ongoing supply constraints and AI demand.

The structural shortage of memory caused Apple to send executives to stay near Samsung and SK hynix facilities to secure long-term supply agreements, a sign of persistent tightness in memory supply chains.

This is not a temporary inventory reshuffle. Memory product lead times and contract pricing both show more sustained imbalances than would be expected in a typical market hiccup.

U.S.-Listed Stocks at the Heart of the Story

Several U.S.-listed companies sit directly at the centre of this memory and storage trend.

Key U.S. Memory & Storage Stocks (Indicative Performance)

Returns are approximate, based on market round-ups and calendar-year performance summaries.

Micron Technology (MU)

Micron is one of the three global memory producers and benefits directly from pricing strength and AI-related demand. It is reported that Micron’s share price rose about 236% in 2025, and management indicated that memory tightness could persist beyond 2026.

Western Digital (WDC) and SanDisk (SNDK)

Western Digital and its spinoff SanDisk are key players in NAND flash and enterprise storage. Reports have highlighted both as major beneficiaries of AI-driven data-centre demand in early 2026 coverage.

Seagate Technology (STX)

Seagate focuses on high-capacity HDDs for data centres. Its enterprise exposure has linked performance closely to AI-led infrastructure expansion.

ETF Exposure to the Theme

For investors seeking diversified exposure rather than individual stocks, semiconductor-focused ETFs provide indirect access.

Semiconductor & Memory-Linked ETFs

Returns are approximate, based on market round-ups and calendar-year performance summaries.

Broader Storage Demand Is Also Rising

AI workloads do not just need memory; they also need persistent high-capacity storage. As enterprise data creation accelerates, both flash and spinning-disk storage demand has grown.

A large industry forecast projects overall data-centre infrastructure spending, including storage and compute, is on track to contribute to a $1 trillion global market by 2030, driven primarily by hyperscale AI build-outs.

What This Means for Indian Investors

For Indian investors accessing U.S. markets via platforms like Appreciate:

• Structural demand from AI data centres has reshaped pricing and supply in global memory markets.
• A highly concentrated supplier base means pricing power can shift rapidly when demand grows faster than capacity.
• U.S.-listed names such as Micron, Western Digital, Seagate, and SanDisk have been among the most directly exposed to this trend.
• Broader semiconductor or tech ETFs with memory and storage exposure may offer diversified access to the theme.

Appreciate combines a wide universe of more than 8,000 U.S. stocks and ETFs with fractional investing and institutional-grade research, making it easier to align global investment themes with individual portfolio objectives.

Conclusion

AI is not just changing software and applications, it is reshaping the physical infrastructure that underpins global digital activity. Memory and storage have moved from relatively cyclical tech segments to the centre of a structural demand shift driven by AI data-centre expansion.

With global memory pricing rising and capacity reallocation tightening supply, a small group of companies controls much of the world’s memory output. For Indian investors looking to participate in this trend via U.S. markets, understanding the underlying economics of memory and storage is now more important than ever.

Visit the new Mint x Appreciate US Markets page — where financial knowledge meets real opportunity.

To know more about investing in US stocks, ETFs, and Mutual Funds, click here.