The market's strongest "caller", one with the surname Chuan, and one with the surname Huang, are both increasing their investments in semiconductors and the next round of technology themes.

Written by: Mike, Frank, MSX.

Since 2025, there have been two men whose "calls" have been the most effective in the market.

One is Huang Renxun; as soon as he stands up at a press conference to talk about GPU, Blackwell, or data centers, the market will reimagine the ceiling of AI; the other is Trump, who, in addition to directly calling out a specific stock, can affect the expectations of an entire industrial chain with his public statements and policy implementations.

Interestingly, just recently, Trump also reported his personal financial status to the government ethics office in accordance with the law, including held stocks, funds, transaction records, and ranges of amounts. Although the disclosure documents cannot prove that every transaction was directly decided by Trump himself and cannot simply be understood as clear buying or selling advice, they at least provide a window for observation:

When a person with the highest policy influence shows obvious directional adjustments in his related accounts, the market will naturally be concerned about what industrial judgments are reflected behind this?

After thorough analysis, MSX found that the most noteworthy aspect of this Q1 disclosure is precisely that Trump's related accounts have begun to conduct intensive trading, with a clear shift toward AI infrastructure, especially significantly cutting back on some older platform technologies and defensive assets, and increasing investment in the supply side of AI infrastructure.

There is no doubt that as the final decision-maker of U.S. policy, the structure of his holdings reflects, to some extent, his judgment on future industrial directions and serves as a window for ordinary investors to understand what the most powerful "smart money" globally is thinking.

1. $220 million in trading volume, over 3,700 transactions

If we first look at the most intuitive data, we will find it is a model of "diligent trading."

According to the disclosed documents, Trump's related accounts completed a total of 3,711 securities transactions in Q1, roughly translating to dozens of transactions per day based on actual trading days; the total trading scale has already exceeded $220 million at the lower limit of the reporting range. Clearly, this is not a quiet account lying still; it is approaching the trading volume of a small to medium-sized hedge fund in a single quarter.

More interestingly, this is very different from Trump's investment style during his first term (2017-2021). At that time, related disclosures showed that he held about 100 individual stocks covering various sectors including finance, healthcare, and industry, resembling a diversified blue-chip portfolio. After entering the White House, he handed over his assets to family and related agencies for management, and his individual stock holdings significantly shrank, with much less active trading than now.

It is worth mentioning that previously, Obama invested funds in treasury bonds and diversified mutual funds, and Biden did not trade stocks at all during his term. Historically, presidents generally choose to divest assets or establish blind trusts to avoid conflicts of interest, while Trump's second-term approach completely broke this convention.

Further analysis reveals a very thematic adjustment in the portfolio.

First, let’s look at where the funds are leaving.

In the first quarter, the largest sell-offs in Trump's related accounts were concentrated in Microsoft, Amazon, and Meta. According to the reporting range, these transactions all touched the highest levels of $5 million to $25 million. These three companies are undoubtedly still core assets in U.S. technology stocks, but they share a common characteristic—they represent the super winners of the previous round of consumer internet, advertising platforms, e-commerce, and cloud services.

Microsoft has software and cloud, Amazon has e-commerce and AWS, and Meta has social networks and advertising systems. They are not lacking in AI stories and are all major players in AI investments, but from a portfolio perspective, these companies have already enjoyed substantial valuation dividends over the past few years. Thus, a significant reduction in holdings may not necessarily indicate bearishness; more accurately, it reflects a reduction in weight on old platform technologies.

It is particularly noteworthy that the disclosure documents did not entirely clear these companies out; some stocks still showed small buying records. This "large sell, small buy" structure resembles an active compression of exposure rather than a complete exit.

Also appearing on the large sell list are dividend-style ETFs like the Vanguard Dividend Appreciation ETF. This indicates that capital is flowing out not only from old tech giants but also includes a portion of defensive and stable assets.

This is crucial. If just selling Microsoft, Amazon, and Meta and then buying another batch of tech stocks, it would only be considered an internal rotation within technology. However, if even defensive ETFs are being reduced, it suggests that the overall risk appetite of the portfolio may be rising, and funds are shifting from stable, old platform assets to more aggressive industrial directions.

So, where is the money going?

The answer is clear—semiconductors, AI hardware, enterprise software, consumer electronics, broad-based indices, as well as some bonds and preferred stocks.

2. From chips to servers, to enterprise software: A systematic coverage of the AI infrastructure chain

If just buying Nvidia, that would only be betting on the AI computing leader, but what is more noteworthy in this disclosure is that Trump's related accounts are not buying a single stock but an entire AI infrastructure chain.

The first layer is semiconductors, including Nvidia, Broadcom, Texas Instruments, Intel, AMD, Micron, and Marvell, which all appear on the buying or increasing shares list. This includes both GPUs and CPUs, both analog chips and storage and interconnect, representing the strongest AI computing leaders commercially as well as stronger policy-oriented representatives of domestic manufacturing in the U.S.; it can be said to cover the entire chain.

Nvidia and Broadcom are self-explanatory. The former is the core target of AI computing, while the latter benefits from trends in custom chips, networking chips, and large cloud vendors' self-developed chips. AMD corresponds to the narrative of GPU alternatives and data center computing, Micron corresponds to storage needs, and Marvell corresponds to interconnection, custom chips, and high-speed data transmission.

Interestingly, Synopsys and Cadence are also on the buying list; these two companies are involved in EDA tools, which are chip design software that ordinary investors might not think of immediately. However, in the semiconductor supply chain, they belong to a very upstream "selling shovels" segment. Every complex chip from design to tape-out almost cannot escape this type of tool, further indicating that this portfolio adjustment is not only chasing the hottest AI leaders but also extending upward along the semiconductor supply chain and bottom-level tools.

The second layer is AI hardware and servers, with Dell being the most sensitive and discussed target among them. The disclosure documents show that Trump's related accounts established a position in DELL within the range of $1 million to $5 million on February 10, and months later, Trump publicly endorsed Dell's hardware products. Subsequently, Dell obtained large government-related contracts, and its stock price also strengthened significantly.

This timeline is sensitive because it involves account buying first, followed by public endorsement, then government procurement, and stock price rise. From a rigorous perspective, the disclosure documents alone cannot prove the causality between the transactions, public statements, and subsequent contracts, but from a market observation angle, such transactions naturally attract attention, as they touch upon highly sensitive nodes such as AI hardware, government procurement, and presidential public statements.

Intel represents another type of sensitivity; unlike Dell, Intel's core is not just commercial logic but also policy logic. The U.S. government has previously decided to make significant equity investments in Intel, and Intel has always been a core target in U.S. domestic semiconductor manufacturing, supply chain security, and industrial policy (see further reading Intel's "life-and-death line" moment: How Chen Liwu settles his legacy and saves himself in front of the ICU?). Against this backdrop, Trump's related accounts purchased INTC multiple times in Q1, which will naturally be interpreted amplified by the market.

Nvidia represents the commercial winner of AI computing, while Intel represents the foundation of domestic manufacturing that the U.S. government wants to support. Their logics differ but point in the same direction: AI infrastructure is no longer just a market theme but is also becoming a direction driven by industrial policy and fiscal resources.

The third layer is enterprise software, including companies like Oracle, ServiceNow, Adobe, and Workday, which also appear on the buying list. Unlike Nvidia, Dell, and Intel, these do not provide computing power and hardware but embed AI directly into enterprise workflows. Oracle corresponds to databases and cloud infrastructure, ServiceNow relates to enterprise process automation, Adobe pertains to creative and marketing productivity, and Workday corresponds to human and financial management systems.

The logic of this layer is also clear: AI cannot merely remain in models and chatbots; it must enter real enterprise budgets, integrate into daily office, customer service, marketing, finance, human resources, development, and data analysis processes. Ultimately, the greatest advantage of enterprise software companies is that they are already in their customers’ workflows. Once AI capabilities become default features of these software, the changes they bring do not just involve new stories but potentially affect renewal rates, pricing power, module upgrades, and customer stickiness (see further reading The "repair" myth of software stocks: After the rebound, are AI agents killers or saviors?).

Therefore, what is truly worth noting in this disclosure is not just which AI hardware companies were bought, but that enterprise software’s AI transformation is emerging as another important clue.

The fourth layer is consumer electronics, for example, Apple received substantial increases in shares and has multiple additional records. Compared to pure AI chips and enterprise software, Apple is more like a representation of AI as an endpoint entry. Whether it can truly drive the AI device cycle remains contested in the market, but in a combination covering AI infrastructure and application endpoints, Apple is undoubtedly a super entry that cannot be overlooked.

Furthermore, the fifth layer involves broad-based indices represented by the S&P 500 ETF, Russell 1000 ETF, and QQQ, which also appear on the large buying list, indicating that this account is not completely detached from the market or unidirectionally betting on a single theme, but actively increasing investments in AI infrastructure and key supply chains while maintaining overall exposure to the U.S. equity market.

Also, the disclosed documents included several bond trades, such as municipal bonds, corporate bonds, high-yield bond ETFs, and preferred stocks, with municipal bonds covering multiple states and corporate bonds including Netflix, Occidental, CoreWeave, etc.

Thus, from the perspective of a portfolio, we can obtain a clear investment self-portrait—on one side, maintaining a foundation and liquidity with broad-based indices, bonds, and preferred stocks; on the other side, enhancing aggressiveness with semiconductors, servers, enterprise software, and AI infrastructure targets.

3. Can we copy the homework?

Upon seeing such disclosures, many people's first reaction might be whether they can follow and buy?

But directly copying homework is not very meaningful, and the reason is simple:

• First, OGE disclosures have a time lag, and by the time ordinary investors see the documents, the trades have already occurred;
• Second, the disclosed amounts are only ranges, not precise figures, such as $1 million to $5 million, $5 million to $25 million, with a vast difference in between, making it hard to accurately gauge the real position weights;
• Third, related accounts may be independently managed by third-party institutions, and the outside world does not know whether each trade is based on active judgment, portfolio rebalancing, or modeled allocation;

Thus, this disclosure is not suitable to serve as a short-term buying signal.

Its real value lies rather in allowing us to see a larger directional change, specifically, the most sensitive "smart money" is transitioning from old platform technology and some defensive assets to the supply side of AI infrastructure, specifically shifting from advertising, e-commerce, and traditional cloud services—core assets of the previous round of the internet—to chips, servers, storage, interconnection, domestic manufacturing, and enterprise software AI transformation.

This direction also has a certain overlap with the current focus of U.S. policy.

After all, domestic semiconductor manufacturing, supply chain security, AI infrastructure, government procurement, and corporate digitization are not purely market stories but directions driven by policy, finance, industry, and capital together, especially for targets like Intel, which has significance not only in terms of performance elasticity but also in reflecting the U.S. desire to regain initiative in advanced manufacturing and the chip supply chain.

This is also the most noteworthy aspect of Trump's related accounts increasing their stake in Intel; it does not necessarily mean Intel is the best chip stock, but it indicates that in the line of AI infrastructure, the market is currently more inclined to pay attention to who is standing at the most concentrated position for policy resources. Similarly, the case of Dell also shows that AI infrastructure is not solely taking place at the GPU level; servers, hardware, government procurement, and enterprise deployment will all become a part of AI capital expenditures landing in the real world.

Therefore, for ordinary investors, what is genuinely worth referencing in this disclosure is not a specific stock but three structural clues.

• AI trading is moving from models and applications to infrastructure: In the past, the market purchased AI more on the anticipation of big models and computing power; now capital is beginning to pay more attention to who can provide chips, servers, storage, networks, packaging, design tools, and enterprise software.
• Semiconductors are no longer just Nvidia: While Nvidia remains the core target, this disclosure shows that capital is also covering Valcom, AMD, Micron, Marvell, Intel, Synopsys, Cadence, and other nodes in the supply chain. As AI infrastructure goes deeper, it is less a single leader's story and more about the revaluation of the entire supply chain;
• Enterprise software AI transformation may be the most underrated segment: Hardware is responsible for building computing power, while enterprise software is responsible for making AI usable. The value of Oracle, ServiceNow, Adobe, and Workday lies not in their ability to tell a completely new AI story but in whether they can embed AI into existing workflows and, through customer stickiness and product upgrades, turn that into revenue.

As for the significant sell-offs in Microsoft, Amazon, and Meta, there is no need to simply interpret that as "these companies will fall." More accurately, it serves as a signal of capital redistribution. After all, once old platform giants have risen significantly, capital will naturally begin to seek assets that are closer to the next round of capital expenditures, more aligned with policy support, and closer to infrastructure development.

All in all, the dividends of the consumer internet era have not yet disappeared, but AI infrastructure, domestic semiconductor manufacturing, and enterprise software AI transformation are indeed accelerating to become the main lines that capital is more willing to pursue in the next phase.

This is also the most noteworthy part of the Q1 portfolio adjustment disclosure from the world's most powerful individual.