The AI investment landscape is being reshaped: Besides the "Big Seven," what other opportunities exist in the semiconductor supply chain?

Since the explosion of ChatGPT at the end of 2022 ignited the artificial intelligence wave, the market’s investment logic around AI has always revolved around the “Magnificent-7,” especially those dominant in cloud computing infrastructure— the “super-large-scale enterprises.” However, the emergence of China’s DeepSeek in early 2025, along with the intense debate over the effectiveness of AI capital expenditures, is quietly changing this landscape. Investors are beginning to realize that the true “gold rush” may not only be with these giants but deeper within the supply chain providing them with “shovels” and “tools.”

From “arms race” concerns to performance validation

In the second half of last year, the market was once filled with worries about AI investment returns. Renowned investor Michael Burry publicly warned that the massive AI capital expenditures by super-large-scale companies might fail to generate expected profits for various reasons, intensifying fears of an AI bubble. At that time, the stocks of the “Magnificent-7” faced pressure, and market sentiment became cautious.

But the earnings season in April this year provided a strong response. Revenue from the cloud computing divisions of these super-large-scale companies continued to surpass expectations, and the robust demand for “computing power” seemed to validate all previous huge investments. From my industry experience, market sentiment often shifts after key data is released. This time was no different; the “hard data” of performance quickly quelled debates over whether capital expenditures were excessive.

The “certainty” dividend of capital expenditure: the semiconductor supply chain explosion

Although discussions about whether AI capital spending will ultimately bring huge profits to these super-large-scale companies are ongoing, a more certain logic has emerged: regardless of who the final winners in AI applications are, this massive capital expenditure will first translate into strong demand for semiconductors and AI-related components.

This judgment directly drove semiconductor-related ETFs to hit record highs in April. From a professional perspective, this is a classic “selling shovels” logic—when a gold rush starts, the people selling shovels often benefit first and most reliably.

Memory chips: the real bottleneck in AI training

In this round of semiconductor rally, memory chip companies performed especially well. U.S. and South Korean giants—SK Hynix, Samsung, SanDisk, Micron Technology, and Western Digital—all saw significant stock price increases. I wrote in March that high-bandwidth memory (HBM) is the true “bottleneck” in AI training. As long as AI’s demand for computing power continues to outpace supply, these companies’ growth momentum will remain unshaken.

Photonics and the broader semiconductor ecosystem

Besides memory chips, photonics companies also performed remarkably. Optical interconnect technology plays a key role in high-speed data transfer within AI data centers, and its importance is being re-evaluated by the market.

Investors have clearly concluded that AI investment opportunities are not limited to the “Magnificent-7.” As of this year, nearly every stock in our tracked “AI-11” semiconductor group has outperformed all but Broadcom among the “Magnificent-7.”

A look into the “AI-11” supply chain: where every dollar flows

Understanding the next AI investment opportunity hinges on seeing how capital flows through this supply chain. Here is my breakdown of the core links in the “AI-11”:

1. Foundry and Lithography (TSMC, ASML)

TSMC provides foundry services for all leading logic chips and is an undisputed industry cornerstone. ASML monopolizes extreme ultraviolet (EUV) lithography equipment, an essential “checkpoint” for manufacturing cutting-edge chips.

2. Logic and Custom Chips (AMD, Broadcom, Intel)

AMD is rapidly gaining market share in AI inference. Broadcom is a key partner in custom ASIC chips for super-large-scale enterprises and dominates network chip markets. Marvell completes the landscape with custom chips, networking, and optical connectivity. Intel is telling its “recovery story” in foundry services and benefits from AI server cycles’ demand for CPUs.

3. Memory Chips (Micron, SK Hynix, Samsung)

These three giants supply the “hard currency” of AI training—high-bandwidth memory. SK Hynix currently leads the global HBM market.

4. Enterprise NAND and Storage (SanDisk, Western Digital)

SanDisk has become a pure beneficiary of enterprise NAND flash and SSDs. Western Digital provides large-capacity mechanical hard drives (HDDs) as a supplement.

Every dollar of capital expenditure on AI infrastructure by super-large-scale enterprises must pass through this complete supply chain before reaching server racks. This explains why the high-tech sector now accounts for a record 55% of U.S. corporate capital expenditure.

The dominance of the “Magnificent-7” remains, but marginal growth is shifting

Undeniably, the “Magnificent-7” still dominate the S&P 500. They account for 30.6% of the index market cap, 25.1% of forward earnings, and 13.7% of forward revenue. The strong fundamentals behind this label still exist.

However, marginal changes are occurring. The forward earnings growth rate of the “Magnificent-7” is currently 25.4%, while the S&P 500 excluding them (the “S&P 493”) stands at 17.9%. This gap has narrowed significantly compared to a year ago. The “S&P 493” is catching up in this growth race. The premium once attributed to the “Magnificent-7” for earnings growth scarcity is becoming less significant as growth spreads more broadly.

I believe the market has already priced in the dominance of the “Magnificent-7.” The marginal capital flowing into the market is shifting toward those sectors capable of extending the AI story beyond these seven companies.

Conclusion: from “betting on winners” to “investing in certainty”

Reviewing the evolution of AI investment logic, a clear main thread emerges: from initially “betting which giant will win” to “investing in the most certain segments of the supply chain.” The certainty of AI capital expenditure has brought unprecedented prosperity to the semiconductor supply chain. For investors, understanding this shift from “demand side” to “supply side” may be key to capturing AI investment opportunities in the coming years. Of course, any investment decision should consider individual risk tolerance; markets always carry uncertainties. But understanding industry logic is the first step toward making wise judgments.