TrendForce significantly raises forecast for the memory market: scale expected to exceed $1.28 trillion by 2027.

```

Agentic AI-driven structural demand expansion is reshaping the global storage market landscape. Market research firm TrendForce has significantly raised its latest industry forecasts, predicting that the global storage market will surpass $1.28 trillion by 2027. The supply gap will be difficult to bridge in the short term, and the momentum for price increases is expected to continue until 2027.

TrendForce has raised its 2026 global storage market forecast from $551.6 billion to $889.3 billion; the 2027 forecast has been revised upward from $842.7 billion to over $1.28 trillion, corresponding to an annual growth rate of about 44%. The magnitude of this adjustment highlights that the construction of AI computing infrastructure is driving storage demand into a new scale.

At the segment level, the DRAM market size forecast for 2026 has been raised to $618.7 billion, a year-on-year increase of 303%, and is expected to further expand to $903.3 billion in 2027, a year-on-year increase of 46%. NAND Flash's 2026 forecast has been raised to $270.6 billion, a year-on-year increase of about 281%, and is expected to approach $379.4 billion in 2027, maintaining an annual growth rate of around 40%.

The core logic behind this round of upward adjustment is the shift in AI development focus from large-scale model training to Agentic AI applications centered on inference, leading to continuous structural demand expansion. On the supply side, the shortage is difficult to resolve in the short term, further strengthening suppliers’ bargaining power in contract pricing.

Agentic AI Reshapes Storage Demand Structure

TrendForce points out that the change in AI development trajectory is the fundamental driver for this round of storage demand expansion. As AI applications shift from one-time large-model training to the continuous inference paradigm of Agentic AI, the characteristics of storage demand are undergoing fundamental changes.

At the DRAM level, inference requests in Agentic AI systems have evolved from one-off queries to continuous iterative loops. At the same time, KV cache capacity grows proportionally with context window expansion—once recalculation is needed, computing costs rise exponentially. Consequently, efficient KV cache management becomes key to AI inference performance, directly driving demand for HBM and DRAM.

In addition, Agentic AI workloads significantly increase CPU requirements for scheduling, data preprocessing, and memory management. TrendForce notes, in next-generation AI server platforms, the CPU-to-GPU ratio has been gradually shifting from the traditional 1:8 to 1:4 or even higher, with Nvidia’s NVL72 rack using a 1:2 configuration. The increase in CPU deployment will in turn boost server DRAM capacity demand, supporting both purchase volumes and contract prices.

DRAM: Supply Constraints and Demand Expansion Drive Both Volume and Price Growth Until 2027

The wafer capacity consumed by HBM production continues to squeeze the available supply of traditional DRAM. In the context of continuously expanding demand, this supply constraint further strengthens suppliers' pricing power in contract negotiations. TrendForce expects price growth momentum to persist through 2027.

TrendForce has raised its 2026 DRAM market forecast to $618.7 billion, an annual growth of about 303%; 2027 revenue is expected to further increase to $903.3 billion, a year-on-year growth of about 46%.

NAND Flash: CSP Capital Expenditures Accelerate, Multi-layer Demand Drives Growth

In the NAND Flash domain, the combined capital expenditures of the nine major global cloud service providers (CSPs) continue to climb rapidly. TrendForce expects the growth rate to reach 79% in 2026, and capital intensity is also expected to rise to 34%. The report notes that this trend reflects a shift in industry strategy from demand-driven capacity expansion to large-scale AI infrastructure investment for long-term competitive advantage.

Agentic AI is also a fundamental driver of NAND demand. AI agents have significantly expanded enterprise-scale usage, with heavy user token consumption up to four times higher than before; the increasing complexity of AI-generated content continues to accelerate token consumption growth.

Facing massive and steadily rising storage demand, the cost of HBM deployment remains relatively high and hard to popularize at scale; traditional HDDs are limited by access speed and power consumption and cannot meet the real-time workload requirements of AI data centers. TrendForce believes that this situation opens up significant growth opportunities for NAND solutions. From SCM SSDs and HBF to SLC/pSLC SSDs, various high-performance SSD technologies are rapidly permeating multiple AI inference, training, and Agentic workloads, becoming new key growth engines.

Against this backdrop, TrendForce has raised its 2026 global NAND Flash market forecast to $270.6 billion, a year-on-year increase of about 281%; the 2027 market size is expected to further expand to nearly $379.4 billion, maintaining an annual growth rate of about 40%.

Risk Disclaimer and Exemption ClauseThe market involves risks and investments should be made cautiously. This article does not constitute personal investment advice nor does it take into account the particular investment objectives, financial situations, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article fit their specific circumstances. Investments made based on this article are at one’s own risk. ```