Trump Just Disclosed Up to $750M in Q1 Stock Trades. The AI Infrastructure Pattern Is Hard to Ignore

Trump Pivoted From Bonds to Big Tech Stocks in Q1 2026

On May 14, the Office of Government Ethics published two new disclosure forms revealing President Trump executed more than 3,700 securities transactions in Q1 2026, with a cumulative value between $220 million and $750 million. For most of his second term, his disclosed activity had been concentrated in corporate and municipal bonds. This is the first quarter showing a sharp pivot into individual stock picking, and the names aren't random.

Two things can be true at once. A sitting president is legally allowed to trade stocks, and prior occupants used managed trusts in similar ways. The cluster of timing across these trades is also the kind of pattern that would draw STOCK Act scrutiny in a congressional case. The more useful question for retail investors isn't whether this is ethical. It's what the trades signal about where policy is pushing capital next.

What Trump Bought and the Timing That's Getting Attention

Among roughly 36 transactions in the $1 million to $5 million bracket, the filing lists Nvidia, Microsoft, Oracle, Broadcom, Amazon, Dell, and Apple, plus at least $740,000 in AMD and a $260,000+ Palantir position. The crypto sleeve includes Coinbase, MARA Holdings, and Strategy Inc. The four largest sells were tech-heavy: between $5 million and $25 million each in Microsoft, Amazon, and Meta on February 10.

The calendar is what's catching attention. A $1 million to $5 million Nvidia buy on February 10 landed seven days before Nvidia and Meta announced their multiyear Blackwell and Rubin GPU partnership, while an earlier January 6 Nvidia buy and a smaller AMD purchase the same day both preceded the Department of Commerce approving H200 and MI325X-equivalent chip sales to China on January 13-14. Oracle shares were accumulated during the same window the administration was helping the company secure the TikTok operating deal, and the Palantir position preceded the February announcement of a five-year, billion-dollar Department of Homeland Security agreement. Trump was fined $200 for missing the disclosure deadline on the Microsoft and Amazon trades by months.

AI Infrastructure Stocks Have Policy Tailwinds Through 2026

The White House defense is that assets sit in a trust managed by Trump's children, with brokers executing trades. That's plausible. It's also impossible to verify from a Form 278-T, which reports only value bands.

Whatever you make of the ethics, the directional signal is hard to argue with. The administration is buying megacap AI names at scale while approving the policies that benefit those exact companies. The China chip story alone is meaningful: Nvidia's H200 China market was previously estimated at roughly $50 billion annually. Greenlighting those sales with a 25% government surcharge unlocks revenue the Street had largely written off. Add the Nvidia-Meta deal and the AMD-Meta 6-gigawatt agreement, and the pattern is unmistakable: U.S. policy is clearing runway for hyperscaler AI spending and chip exports at the same time senior insiders are accumulating those names.

Where the Second-Derivative Trade Lives: Neocloud and AI Compute

If the megacap chips and hyperscalers are already crowded and priced for substantial growth, the compute layer one rung down the stack offers the more interesting setup. Three neocloud names sit directly downstream of every dollar of hyperscaler capex: CoreWeave, the largest GPU-as-a-service neocloud with multi-billion-dollar hyperscaler contracts already locked in; Nebius, with a revenue growth profile that looks more like an early-stage hyperscaler than a traditional cloud provider; and IREN, which sits at the intersection of Bitcoin mining and AI compute with administration policy favoring both buildouts.

The base case scenario (roughly 55-60% probability) is that AI infrastructure policy support stays intact through 2026, hyperscaler capex grows another year, and neocloud revenue compounds aggressively from a small base. The bull case (20-25%) sees the China chip market unlock faster than expected. The bear case (15-20%) is a credit-market wobble or sudden regulatory reversal that re-rates everything tied to AI capex.

How to Position Without Predicting Politics

Most retail investors already own the megacap AI names through index funds, since the Magnificent Seven now represent roughly 25% of S&P 500 earnings. For higher conviction, direct positions in Nvidia, AMD, or Broadcom remain the cleanest plays. For higher beta on the same thesis, the neocloud names offer levered exposure at smaller market caps. Position sizing matters more than picking the perfect ticker, since the velocity of policy reversals in 2026 has been higher than in any year of the past decade.

Key Takeaways

• Trump's Q1 stock buys cluster around favorable regulatory decisions in same names
• AI infrastructure has clear policy tailwinds heading into late 2026
• Neocloud names offer second-derivative exposure to hyperscaler AI capex

---

Disclaimer: This article is for informational purposes only and should not be considered financial advice. Stock investing involves significant risk, including potential loss of principal. Always conduct your own research and consult with a qualified financial advisor before making investment decisions.