The core of the last AI market rally was highly concentrated, with trading activity revolving almost exclusively around two main themes: computing power and storage. In particular, the market focused on NVIDIA and HBM. Following Micron’s earnings report, which reinforced demand expectations, and SK Hynix’s continued leadership in HBM technology, the storage sector became one of the most crowded areas for capital inflows. However, as we approached late June, a clear shift emerged: while HBM remained in a strong uptrend, its rate of ascent slowed, volatility increased, and investors began seeking new sources of returns. This shift doesn’t signal the end of the AI rally. Instead, it’s a classic sign that the market is moving from a "single-theme-driven" phase to a "sector rotation across the value chain."
The Nature of HBM’s High-Level Volatility: From Expectation-Driven to Realization Phase
HBM’s recent volatility isn’t due to a sudden deterioration in fundamentals, but rather a shift in market trading dynamics. Previously, HBM’s rally was fueled primarily by the "supply-demand gap + AI expansion expectations"—a classic expectation-driven phase. As Micron and SK Hynix continued to signal capacity expansion and secure long-term orders, some of these expectations were priced in ahead of time, naturally leading the market into a rebalancing phase.
At the same time, the storage industry is inherently cyclical. Even with AI-driven demand, price elasticity typically follows a "rapid rise—high-level stability—structural divergence" pattern. The market now appears to be in the early stage of the third phase, where prices no longer rise unilaterally but instead fluctuate at elevated levels, awaiting new demand catalysts or narrative expansion.
First Wave of Capital Rotation: Data Centers and Cloud Infrastructure
As HBM enters a consolidation phase, capital often flows first to its upstream demand drivers—data centers and cloud infrastructure. The logic is straightforward: HBM demand is driven by GPUs, and GPU demand is in turn driven by data center expansion.
So, when the storage segment’s momentum slows, capital naturally returns to the "source of demand." At this stage, the market’s focus shifts from individual chips to overall AI infrastructure capital expenditures, including server expansion by cloud providers, growing AI inference needs, and accelerated enterprise AI deployments. These assets are characterized by longer cycles and higher certainty, making them more stable options during sector rotations.
Second Wave of Rotation: Repricing Leading Computing Power Assets
During storage sector volatility, another common path is for capital to rotate back into core computing power assets. Although leaders like NVIDIA have already seen significant gains, their central role in the AI ecosystem remains unchanged.
HBM’s fundamental role is to "enhance GPU efficiency," not to replace GPUs. As the storage segment consolidates, the market often refocuses on the long-term demand curve for computing power. Particularly as AI moves from training to inference, demand for computing power continues to grow—though the pace is shifting from explosive to steady.
This phase is marked by a realignment of valuations and growth, rather than the start of another one-sided rally.
Third Wave of Diffusion: Catch-Up in AI Applications and Software Layer
Once upstream infrastructure enters consolidation, capital typically spreads downstream into AI applications and the software layer. This segment is characterized by lagging valuations, broader narrative potential, but longer performance realization cycles.
AI applications include enterprise AI tools, industry-specific vertical models, AI agents, and various software service platforms. These assets often benefited less in the initial phase but can see catch-up rallies during sector rotation. The market’s logic shifts from "infrastructure certainty" to "application-side flexibility."
Thus, capital behavior in this stage is more theme-driven and focused on expectation recovery, rather than rapid fundamental validation.
Fourth Wave of Extension: Optical Communication and Energy Infrastructure
Beyond computing power and applications, another area gaining market attention is the peripheral systems of AI infrastructure, including optical communication, high-speed interconnects, and energy systems.
As AI clusters scale up, the efficiency of chip interconnections and data transmission becomes increasingly critical, making optical modules and high-speed interconnect technologies new bottlenecks. At the same time, the rising power demands of large-scale data centers are turning energy infrastructure into a long-overlooked but increasingly important variable.
This segment is characterized by "latent demand," which often gets fully priced in only in the later stages of industry expansion.
Structural Shifts in the AI Value Chain: From "Single-Point Breakouts" to "Chain-Based Rotation"
The most significant change in the current AI market isn’t simply price movement, but a transformation in market structure. Previously, the market’s core pattern was single-point breakouts—such as rallies led by GPUs or HBM. Now, the structure is shifting toward chain-based rotation, with capital spreading up and down the value chain.
This shift means that, on one hand, excess returns from single assets are decreasing; on the other, the overall extensibility of the value chain is increasing, potentially prolonging the cycle. In other words, the AI market is transitioning from a "strong, concentrated trend" to a "multi-stage rotational rally."
Gate Stock Trading: Enhancing Trading Efficiency Amid Cross-Market Rotation
As AI enters a rotation phase, a practical challenge is the increasingly dispersed market landscape. US, Korean, and Hong Kong stocks each represent different segments of the value chain—such as storage, computing power, and applications—making cross-market tracking more essential.
Against this backdrop, Gate Stock Trading offers 24/7 trading for US, Hong Kong, and Korean stocks, enabling investors to flexibly reallocate assets across markets and more promptly participate in the AI value chain rotation. For event-driven tech rallies, this trading mechanism significantly enhances cross-market responsiveness.
Conclusion: The AI Rally Enters a More Complex but Longer-Lasting Phase
HBM’s high-level volatility doesn’t signal the end of the AI rally. Instead, it marks the market’s transition from a single-theme focus to a phase of structural rotation. Capital is spreading from storage to computing power, data centers, applications, and peripheral infrastructure. This shift makes the AI rally more complex, but potentially more durable.
Going forward, AI investment strategies will shift from seeking a single main theme to understanding how the value chain rotates and how capital reallocates across different segments.
FAQs
1. Why did HBM shift from a strong uptrend to volatility?
Mainly because earlier expectations were quickly realized, and the industry is now entering a supply-demand rebalancing phase.
2. Why is capital flowing out of HBM?
After concentrated gains were realized, investors began searching for new sources of growth potential.
3. What are the most likely beneficiaries in the next phase?
Data centers, leading computing power assets, AI applications, and optical communication infrastructure.
4. Has the AI rally ended?
No—it has shifted from a single-theme focus to a value chain rotation phase.
5. What is the significance of Gate’s 24/7 stock trading?
It enhances cross-market participation efficiency, allowing investors to more flexibly capture opportunities at different stages of the AI value chain.
