Since 2025, two men have been most effective at “recommending stocks” in the market. One is Jensen Huang, as long as he stands on the stage at a launch event, talking about GPUs, Blackwell, and data centers, the market will reimagine the ceiling of AI; the other is Trump, in addition to directly touting a particular stock, his public statements and policy implementation will affect the expectations of an entire industry chain.

Recently, Trump legally declared his personal financial situation to the government ethics office, including stocks, funds, transaction records, and amount ranges held. Although the disclosure documents cannot prove that every transaction was personally decided by Trump, nor can it be simply understood as a clear buy or sell recommendation, it provides an observation window: when a person with the most policy influence sees obvious directional adjustments in his related accounts, the market will of course care about what kind of industrial judgment is reflected behind this?

After in-depth analysis, MSX found that the most noteworthy part of this Q1 disclosure is that Trump’s related accounts began to trade intensively, and the direction clearly shifted to AI infrastructure, especially the large-scale compression of some old platform technology and defensive assets, and increased investment in the supply side of AI infrastructure. There is no doubt that as the ultimate decision-maker of American policy, his holding structure, to some extent, reflects his judgment on the future direction of the industry.

I. $220.00M Trading Volume, Over 3700 Transactions

If you look at the most intuitive data first, you will find that it is a model of “diligent trading.” According to the disclosure documents, Trump’s related accounts completed a total of 3711 securities transactions in Q1, and the cumulative transaction scale has exceeded $220.00M based on the lower limit of the declared range. This is clearly not a quiet, static account, and is already close to the trading volume of a small and medium-sized hedge fund in a single quarter.

This is very different from Trump’s investment style during his first term (2017-2021), when he held about 100 individual stocks, which overall looked more like a diversified blue-chip portfolio. It is worth mentioning that Obama previously invested funds in treasury bills and diversified mutual funds, while Biden did not trade stocks at all during his tenure. Previous presidents generally chose to divest assets or establish blind trusts to avoid conflicts of interest, while Trump’s approach in his second term completely broke this convention.

If you break it down further, you can find a very thematic portfolio adjustment. In the first quarter, the largest sales in Trump’s related accounts were concentrated in Microsoft, Amazon, and Meta. These companies are undoubtedly still core assets in American technology stocks, but they represent the super winners of the previous round of the consumer internet, advertising platform, e-commerce, and cloud service era. A large reduction in holdings does not necessarily mean being bearish, but more accurately, it is reducing the position weight of old platform technology.

Also appearing on the list of large sales is the Vanguard Dividend Appreciation ETF, a dividend-style ETF. This shows that funds are flowing out not only from old technology giants, but also from some defensive and stable assets. This indicates that the overall risk preference of the portfolio may be rising, and funds are shifting from stable and old platform assets to more aggressive industrial directions.

II. From Chips to Servers to Enterprise Software: AI Infrastructure Chain Systematically Covered

If you only buy Nvidia, you are only betting on the leading AI computing power company, but what is more noteworthy in this disclosure is that Trump’s related accounts are not buying a single target, but an entire AI infrastructure chain. The first layer is semiconductors, including Nvidia, Broadcom, Texas Instruments, Intel, AMD, Micron, Marvell, etc., covering GPUs, CPUs, analog chips, storage and interconnect, as well as EDA tool vendors Synopsys and Cadence.

The second layer is AI hardware and servers, with Dell being the most sensitive target. Disclosure documents show that Trump’s related accounts established a DELL position in the range of $1.00M to $5.00M on February 10. A few months later, Trump publicly endorsed Dell hardware products. Intel is another kind of sensitivity. As a core target of American domestic manufacturing, Trump’s related accounts repeatedly bought INTC in the first quarter, which was interpreted by the market as a direction jointly promoted by industrial policy and financial resources.

The third layer is enterprise software, including Oracle, ServiceNow, Adobe, Workday, etc. The logic is that AI will eventually enter real enterprise budgets and daily workflows. The fourth layer is consumer electronics, such as Apple, which received a large increase in holdings, as an AI terminal entry point. The fifth layer is broad-based indexes such as the S&P 500 ETF, Russell 1000 ETF, QQQ, as well as municipal bonds, corporate bonds and other bond assets.

III. Can You Copy the Homework?

Seeing this kind of disclosure, many people’s first reaction may be whether they can follow the buying? But directly copying the homework is not very meaningful. First, OGE disclosure has a time lag; second, the disclosed amount is only a range, and it is difficult to judge the true position weight based on this; finally, the related accounts may be independently managed by a third-party institution. Therefore, this disclosure is not suitable as a short-term buying and selling signal. Its real value is to let us see that “smart money” is shifting from old platform technology to the supply side of AI infrastructure.

For ordinary investors, what is really worth learning from this disclosure are three structural clues: AI trading is shifting from models and applications to infrastructure; semiconductors are no longer just Nvidia, but the revaluation of the entire supply chain; and the AI-ization of enterprise software may be a more easily underestimated link.

As for the large reduction in holdings of Microsoft, Amazon, and Meta, this is more of a signal of capital reallocation. No matter what, the dividend of the consumer internet era has not disappeared, but AI infrastructure, semiconductor localization, and enterprise software AI-ization are indeed accelerating to become the main lines that funds are more willing to pursue in the next stage.

[MSX Maitong]