Trump's Q1 2026 portfolio: what the OGE filings actually disclosed

The US Office of Government Ethics published two Form 278-T Periodic Transaction Reports for President Donald J. Trump on May 14 and May 15, 2026. Together they document 3,711 securities transactions executed January 6 → March 30, 2026, with a cumulative reported value of $220M–$750M (OGE discloses in bands, not exact dollar amounts). This is the largest single-quarter trading footprint disclosed by any sitting US president, by a margin that isn't close.

We pulled every ticker flagged in news coverage of the filings, cross-referenced against the 58-position holdings reconstruction published by @Speculator_io, and mapped the result onto the QuantAbundancia bubble taxonomy. Live data: /portfolios/donald-trump.

The headline: AI-supercycle concentration, executed Feb 10

The single most informative date in the filings is February 10, 2026. On that one trading day the account:

• Sold $MSFT, $AMZN, $META — each in the $5M–$25M band
• Bought $DELL in the $1M–$5M band (Dell hardware was publicly endorsed by the President at a White House event in early May)
• Bought additional $NVDA, $AVGO, $AAPL, $MSFT in the $1M–$5M band (yes, MSFT both sides — the larger leg was the sell)

The rotation is unambiguous: out of three hyperscaler megacaps that QuantAbundancia's residualized correlation work tags as the failed AI bubble (Hyperscalers co-move with SPY beta, not with each other as a thematic cluster), into single-name semiconductor and AI-hardware exposure that maps directly onto the validated bubbles.

The Q1 buy cluster — semis, AI hardware, EDA, crypto

By ticket count, the most-bought single names in the disclosed account were:

| Ticker | Buys | Sells | Theme | | --- | --- | --- | --- | | $ORCL | 11 | 0 | Cloud DB + OCI AI infrastructure | | $COST | 11 | 0 | Defensive megacap retail | | $AAPL | 8 | 1 | Megacap consumer hardware | | $NVDA | 7 | 2 | AI-compute anchor | | $CDNS | 7 | 0 | EDA duopoly | | $SNPS | 6 | 2 | EDA duopoly | | $ADBE | 5 | 3 | Enterprise creative SaaS (-32% YTD) | | $WDAY | 5 | 3 | HCM SaaS (-45% YTD) | | $NOW | 4 | 2 | Workflow SaaS (-41% YTD) | | $AMD | 4 | 4 | AI-compute alt-name | | $DIS | 7 | 0 | Media | | $SBUX | 4 | 0 | Consumer staples | | $WFC | 4 | 0 | Money-center bank | | $GS | 4 | 0 | Money-center bank |

The semiconductor + EDA pattern is unusually clean. $NVDA (7 buys) + $AVGO + $TXN (3 buys) + $KLAC + $AMKR + $CDNS (7 buys) + $SNPS (6 buys) + $MU + $INTC is a basket that covers compute, networking, analog/embedded, fab equipment, advanced packaging, EDA software, memory, and the IDM-comeback bet — the full Semi Equipment + Compute Capacity stack.

The AI-hardware leg pairs $DELL (+93% YTD, the best-performing position in the book) with $JBL Jabil (3 buy tickets, +53% YTD). Jabil is the contract assembler for AI server racks; Dell is the system integrator selling them. The two are a vertically-stacked thesis.

Enterprise SaaS appears as a deliberate buy-the-dip basket: $ADBE (-32% YTD, net buyer), $NOW (-41% YTD, net buyer), $WDAY (-45% YTD, net buyer), $CRM (-36% YTD), $FIS (-36% YTD). All four SaaS names are deeply negative on the year and all four show buy tickets in the PTRs. The account is taking the other side of the AI-replaces-SaaS narrative that drove the drawdown.

The crypto-linked equity basket: $COIN Coinbase (9 confirmed buys), $MARA MARA Holdings, $MSTR Strategy (formerly MicroStrategy). All three are flagged in CryptoTimes' coverage of the May 15 PTR. The Strategy buy is particularly notable — it puts the President's disclosed account onto Michael Saylor's bitcoin-treasury vehicle while in office.

The Q1 sell cluster — Mag-7 trim, broad-market ETFs liquidated

The disclosed account sold three hyperscaler megacaps in size on February 10:

| Ticker | Sell size | YTD | | --- | --- | --- | | $MSFT | $5M–$25M band | -15% | | $AMZN | $5M–$25M band | -16% | | $META | $5M–$25M band | (not in image — fully exited?) |

$MSFT and $AMZN are still surfaced in the speculator_io image (trimmed but held); $META does not appear in the 56-name reconstruction, suggesting the position was sold to zero. The three trades on the same day at the same band suggest a deliberate, sized rotation, not opportunistic profit-taking.

The broad-market ETF unwind is the second-quietest tell in the filings:

• $VIG Vanguard Dividend Appreciation — sold ~$15M on January 12
• $SPY — net seller (mixed activity)
• $VOO — 2 sales, zero buys
• $IVV — sold $3M on January 23
• $XLC Communication Services SPDR — sold $3M on March 5
• $VYM Vanguard High Dividend Yield — sold $1M on January 23

The pattern is a coordinated exit from broad-market and dividend-style beta in the first three weeks of January, then redeployment into the single-name semiconductor + AI-hardware + crypto-linked basket through February and March. Reverse-engineered, the account moved from "S&P 500 + dividend tilt" to "concentrated AI-supercycle + crypto" inside a single quarter.

What the QA bubble taxonomy says

Mapping the disclosed Q1 activity onto QuantAbundancia's 12 validated AI bubbles:

• Semi Equipment — direct hits: KLAC, CDNS, SNPS, AMAT (buy), LRCX (buy). The 0.82-residualized-correlation cluster QA flags as the tightest validated bubble in the entire taxonomy. The disclosed account owns it as a basket.
• Compute Capacity — NVDA, AMD, AVGO. All net-bought.
• Memory — MU is named in CNBC's coverage as a $500K–$1M band buy. SanDisk does not appear in the speculator_io image but may be in the raw PTRs.
• Datacenter Power — ETN, XEL (carryover from prior 278). BE (Bloom Energy) appears in CNBC's coverage as a $500K–$1M band buy. The same theme Druckenmiller doubled on in Q1 (VST +114%) shows up here in smaller size.

The notable absence: Networking & Optical direct exposure (LITE, COHR, ALAB, CRDO) doesn't surface in either the news coverage or the image, despite being one of the cleanest Q1 themes in the institutional 13Fs. AVGO is the only adjacent name, and it covers networking only as a small slice of its custom-silicon revenue.

How this compares to the Q1 13F tape

The 13F filings published last week split the institutional tape into three clusters: AI-bottleneck believers (Tiger Global, Duquesne), megacap-compounder buyers (Ackman, Buffett, TCI), and real-economy + deep-value outliers (Gates Foundation, Greenlight).

The disclosed Trump account does not fit any of the three cleanly. It's closest to Duquesne on the AI-supercycle direction (rotation out of consensus megacaps into bottleneck names) but with more breadth — 58 holdings vs Duquesne's 10 in his top book, and a heavy enterprise-SaaS buy-the-dip overlay Duquesne doesn't carry. It's closest to Tiger Global on the semi-equipment basket, but Tiger also owns the optical names (LRCX, AMAT, plus ALAB-adjacent custom-silicon plays) and Trump's account doesn't.

The two cleanest reads:

1. Same end-state as Druckenmiller, broader path. Druckenmiller sold NVDA and concentrated into memory + power. The disclosed Trump account bought NVDA and concentrated into semis + AI hardware + crypto. Different roads, same destination — both books are rotating capital out of consensus AI megacaps and into more constrained-resource AI bets.
2. Faster cadence than any institutional book here. Tiger and Duquesne file 13Fs quarterly and publish ~30-line positions. The disclosed Trump account filed two PTRs in a week covering 3,711 trades. The frequency is closer to a CTA or systematic-equity book than to a long-only institutional portfolio.

Takeaway — observational, not predictive

This article surfaces what the filings actually disclose. We're not predicting whether the discretionary-managed account will outperform; the Pelosi STOCK Act book showed last year that political-figure portfolios can compound aggressively, but the management overlay (and the legal scrutiny it invites) introduces variance other portfolios don't have.

For traders, the cleanest read is: another high-profile discretionary book added size to the same AI-bottleneck cluster QA's bubble taxonomy already validates. The Druckenmiller + Tiger + Aschenbrenner + Trump-disclosed-account quartet now share Semi Equipment + Compute Capacity as a basket. That's four institutional-scale rebalancing flows pointing the same direction.

Live per-position data, updates when fresh PTRs file, and the full 58-name holdings reconstruction at /portfolios/donald-trump. Methodology context: Correlation over narrative and Residualized correlation.