Stanley Druckenmiller — 13F Portfolio Analysis (2025)
Entity: Duquesne Family Office LLC · CIK 0001536411
Coverage: Q1–Q4 2025 (Form 13F filings, period endings Mar / Jun / Sep / Dec 2025)
Data source: SEC EDGAR Form 13F-HR via OpenBB SDK (obb.equity.ownership.form_13f)
Disclaimer: Form 13F discloses long equity and call option positions only.
Short positions, futures, currencies, bonds, and macro derivatives are not reported.
The 45-day filing lag means Q4 2025 data reflects a Dec 31 2025 snapshot — not current holdings.
This report is produced for financial education purposes only and is not investment advice.
Data Sources & Citations
| # | Source | URL |
|---|---|---|
| [1] | SEC EDGAR Form 13F-HR (primary) | https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001536411&type=13F |
| [2] | 13Radar — Q4 2025 ETF Pivot | https://13radar.com/guru/stanley-druckenmiller |
| [3] | WhaleWisdom — Duquesne 13F | https://whalewisdom.com/filer/duquesne-family-office-llc |
| [4] | 13f.info — Duquesne Family Office | https://13f.info/manager/0001536411-duquesne-family-office-llc |
| [5] | Motley Fool — Druckenmiller Natera Thesis | https://www.fool.com/investing/how-to-invest/famous-investors/duquesne-family-office/ |
| [6] | OpenBB SDK — obb.equity.ownership.form_13f |
https://docs.openbb.co |
| [7] | StockCircle — Druckenmiller Nov 2025 | https://stockcircle.com/portfolio/stanley-druckenmiller |
1. Portfolio Overview
Duquesne Family Office LLC’s disclosed AUM grew 46.9% during 2025,
from $3,060M (Q1) to $4,494M (Q4). [1][6]
| Quarter | AUM ($M) | Equity Positions | HHI | Top-1 Wt | Top-10 Wt |
|---|---|---|---|---|---|
| Q1 2025 (Mar) | $3,060M | 49 | 0.0546 | 15.7% | 59.8% |
| Q2 2025 (Jun) | $4,071M | 65 | 0.0403 | 12.8% | 50.5% |
| Q3 2025 (Sep) | $4,062M | 62 | 0.0449 | 12.7% | 53.9% |
| Q4 2025 (Dec) | $4,494M | 58 | 0.0422 | 12.8% | 52.0% |
Source: SEC EDGAR Form 13F via OpenBB [1][6]. Cross-checked against 13f.info [4] and WhaleWisdom [3].
Key findings
- The sharpest AUM expansion was Q2 (+$1.0B), driven by 31 new position initiations.
- Q3 was flat ($4,062M vs $4,071M) — a consolidation phase.
- Q4 set a new peak as a large ETF bloc pushed AUM to $4,494M. [2]
Educational note — HHI: The Herfindahl-Hirschman Index measures concentration. A fully
diversified 100-stock equal-weight fund scores ~0.01; a single-stock fund scores 1.0.
Druckenmiller’s HHI of 0.040–0.055 sits in the
concentrated-but-not-extreme zone — active conviction without reckless concentration.
2. Concentration & Structure
The book has four structural layers:
| Layer | Characteristics | Q4 Example |
|---|---|---|
| Anchor | 1 name, 12–16% of AUM, held continuously | Natera (NTRA) 12.8% = $575M |
| Thematic sleeve | 30–40 names, ~50–60% of AUM, 60%+ quarterly churn | Insmed, TSMC, Coupang, Sea Ltd |
| Macro overlay | 3–10 ETF/index positions, grew to 21% in Q4 | SPY, IWM, EWZ, RSP, XLF |
| Moonshots | ~10 names at 0.5–1% each | Newamsterdam Pharma, Cogent Bio |
Educational note — conviction sizing: Druckenmiller’s anchor sits at ~13% of book —
not 30–40% as a retail investor might weight their “best idea”. Even the world’s most
conviction-driven macro trader caps his largest disclosed position at ~16%. This preserves
optionality and prevents one thesis from destroying the portfolio.
3. Core Positions Held All 4 Quarters
Only 15 positions survived every quarter. Share count delta — not dollar value — reveals true
conviction direction, because it removes the price-appreciation effect.
Verified from raw SEC filings [1][6]:
- Natera (NTRA): Q1 $481M -> Q4 $575M.
Shares -26.2% — he trimmed while dollar value grew.
Price appreciation did the work, not new buying. [1][2][5]
- Newamsterdam Pharma (NAMS): Q1 $16M -> Q4 $108M.
Shares +284% — systematic accumulation, not a lucky spike. [1][6]
Educational note — share count vs dollar value: A position can double from $50M to $100M
in two very different ways: (1) the stock doubled — no new conviction required; or
(2) you bought more shares — an active conviction signal. Share count is the honest metric.
Dollar value is noisy because it conflates price moves with portfolio decisions.
4. Thematic Rotation
| Theme | Q1 | Q2 | Q3 | Q4 | Direction |
|---|---|---|---|---|---|
| Biotech / Pharma | 31.5% | 29.0% | 35.1% | 27.7% | down reducing |
| AI / Semis / Tech | 14.6% | 20.4% | 17.3% | 20.3% | up building |
| Macro / ETF | 1.2% | 5.5% | 6.5% | 20.9% | up Q4 surge |
| Emerging Markets | 14.2% | 10.7% | 9.7% | 10.8% | down gradual |
| Industrials / Defense | 9.6% | 9.3% | 8.7% | 6.0% | down reducing |
| Airlines / Travel | 5.9% | 4.5% | 0.7% | 2.6% | down mostly out |
| Energy / Resources | 6.3% | 3.7% | 2.8% | 3.2% | down reduction |
| Financials | 3.6% | 3.2% | 3.2% | 0.5% | down fully exited |
| Consumer / Retail | 1.3% | 2.7% | 6.3% | 4.7% | up then partial out |
Source: SEC EDGAR Form 13F via OpenBB; thematic classification applied by this analysis [1][6]
Q4 headline shift: ~$940M entered as ETF positions (SPY/IWM calls, EWZ calls + equity,
RSP equity, XLF equity, iShares EM equity) — a top-down broadening bet signalling he expected
the bull market to rotate beyond mega-cap tech into small-caps, EM, and cyclicals. [2][3]
Educational note — theme exits vs stock exits: Druckenmiller doesn’t rotate within themes
— he exits themes entirely. Financials went from 3.6% to 0.5% in a single year: not a trim,
an abandonment. This is portfolio-level thinking — bet on the theme, not the individual stock;
when the thesis changes, exit the whole bloc.
5. Options Usage
Druckenmiller uses options exclusively as upside amplifiers, never as hedges.
Verified: 0 put positions across all 4 quarters (zero). [1][6]
| Quarter | Position | Type | Value | Education Note |
|---|---|---|---|---|
| Q1 | Delta Air Lines (DAL) | CALL | $41M | Airlines thesis amplification |
| Q1 | Seagate Technology (STX) | CALL | $21M | Storage / AI adjacency |
| Q1 | United Airlines (UAL) | CALL | $16M | Paired trade with DAL |
| Q2 | iShares Russell 2000 (IWM) | CALL | $72M | 335K shares — small-cap broadening |
| Q2 | SPDR S&P 500 (SPY) | CALL | $56M | 90K shares — index macro overlay |
| Q2 | Natera (NTRA) | CALL | $19M | Top-conviction double-up |
| Q2 | Insmed (INSM) | CALL | $10M | Biotech amplification |
| Q3 | SPDR S&P 500 (SPY) | CALL | $60M | Same 90K shares as Q2 |
| Q3 | iShares Russell 2000 (IWM) | CALL | $55M | Trimmed to 226K from Q2’s 335K |
| Q3 | Natera (NTRA) | CALL | $16M | Maintained |
| Q4 | iShares Brazil (EWZ) | CALL | $134M | Largest options position of the year |
| Q4 | SPDR S&P 500 (SPY) | CALL | $61M | Same 90K shares — held 3 quarters unchanged |
| Q4 | iShares Russell 2000 (IWM) | CALL | $56M | Same 226K shares as Q3 |
| Q4 | Amazon (AMZN) | CALL | $23M | Re-entered via options before equity |
Correction from initial analysis: The original report stated IWM calls were held at
“226K shares from Q2 through Q4.” SEC data shows Q2 had 335K shares, trimmed to 226K
from Q3 onward. SPY was correctly 90K unchanged across all three quarters.Educational note — structured overlays vs trades: The SPY call (90K shares) sat unchanged
across Q2–Q4 — bought once, carried forward. This is a structural overlay on index upside,
not an actively managed position. The Brazil call entering at $134M in Q4 is the active signal
— a new, high-conviction EM re-rate bet. [1][6]
6. Capital Deployment & Turnover
| Transition | New In | Exited | Held | Churn % |
|---|---|---|---|---|
| Q_prev -> Q2 2025 (Jun) | 31 | 15 | 34 | 57.5% |
| Q_prev -> Q3 2025 (Sep) | 29 | 32 | 33 | 64.9% |
| Q_prev -> Q4 2025 (Dec) | 26 | 30 | 32 | 63.6% |
Source: SEC EDGAR Form 13F position-level comparison [1][6]
~60–65% of positions turn over every quarter — extremely high for a $4B+ fund.
Most institutional funds of this size have annual turnover of 20–40%, not quarterly.
Educational note — high conviction does not equal low turnover: Many investors equate
long-term conviction with low turnover (a la Warren Buffett). Druckenmiller demonstrates a
different model: high conviction with a defined catalyst horizon. When the catalyst plays
out or fails, the position goes — regardless of tax considerations or ego. This isn’t
indecision; it’s discipline about catalysts vs noise.
7. Geographic Exposure
| Quarter | US Domestic | International | Key Drivers |
|---|---|---|---|
| Q1 2025 | 65% | 35% | Argentina cluster (YPF, Banco Macro, Galicia) |
| Q2 2025 | 71% | 29% | New US tech positions dominate new capital |
| Q3 2025 | 70% | 30% | EM positions stable; Verona / Warner added |
| Q4 2025 | 77% | 23% | SPY / IWM ETF bloc overwhelmingly US-focused |
Source: SEC EDGAR Form 13F; geographic classification based on primary listing [1][6]
Educational note — ADR vs underlying: “US exposure” in 13F terms is a proxy, not a true
measure. A US-listed ADR (e.g. TSMC’s NYSE ADR) is classified as “US-listed” even though the
underlying business is Taiwanese. True geographic exposure requires looking through to business
domicile and revenue geography.
8. The Druckenmiller Method — Synthesised
Six operating patterns emerge from four quarters of verified 13F data:
Pattern 1 — One Massive Anchor + Calls on Top
Natera held at 13–16% of book all year. Call options added in Q2–Q3 to amplify upside
without increasing absolute downside. Shares trimmed -26.2% while dollar
value grew $481M -> $575M. [1][6]
Pattern 2 — Rotating Thematic Sleeve (60%+ Quarterly Churn)
Catalyst-driven, not sentiment-driven. Eli Lilly: in Q1, out Q3 — GLP-1 trade fully priced.
Microsoft: in Q2, out Q3 — one-quarter tactical. No ego, no anchoring, just catalysts.
Pattern 3 — Macro Overlay via Liquid ETFs
Top-down views expressed through index instruments, not individual equities. In Q4 this
grew to 20.9% of the disclosed portfolio — a loud signal about broadening risk appetite.
SPY calls held unchanged three consecutive quarters: a structured bet, not a trade. [2][3]
Pattern 4 — Small-Cap Moonshots (0.5–1% sizing)
Newamsterdam Pharma: $16M -> $108M, shares +284%.
Many will go to zero; a few generate 5–10x. The sizing discipline limits losses while
preserving optionality on outliers. [1][6]
Pattern 5 — EM as a Multi-Vehicle Structural Belief
Emerging markets expressed simultaneously through direct equities (Coupang, Sea Ltd,
MercadoLibre, Nu Holdings), country ETFs (iShares Brazil, Global X Argentina), and ADRs.
The thesis is held at portfolio level — when Argentina disappoints, he exits all vehicles.
Pattern 6 — Zero Ego on Re-Entry
Entegris: bought Q2, sold Q3, re-bought Q4. Microsoft: bought Q2, sold Q3.
The data shows a mind unburdened by the psychological cost of “admitting you were wrong.”
The question is always: “Is the thesis valid today?” — never “Was I wrong before?”
9. Limitations of 13F Analysis
Warning: This section is critical for sound interpretation of this report.
| Limitation | Impact | Magnitude |
|---|---|---|
| Long-only disclosure | Shorts, futures, bonds, currencies not shown | Could be 2–5x the disclosed longs |
| 45-day filing lag | Q4 data filed Feb 17 2026 — reflects Dec 31 snapshot | Holdings may be materially different today |
| No leverage data | Dollar values exclude margin; real exposure unknown | Could be 1.5–3x levered |
| No cost basis | Cannot compute unrealised P&L or entry timing | Performance attribution impossible |
| Calls shown, puts hidden | Protective puts (hedges) not visible in 13F | Hedging strategy fully opaque |
| ADR vs underlying | TSMC ADR != TSMC equity; pricing may diverge | Small distortion in values |
The 13F is a keyhole into one room of a very large house.
Druckenmiller is famous for macro bets in currencies and futures that dwarf his equity book.
The 13F likely captures 20–50% of his total position exposure at best.
Analysis generated using OpenBB + SEC EDGAR. Generated: 2026-04-06
Cross-referenced: 13Radar [2], WhaleWisdom [3], 13f.info [4], Motley Fool [5]
For financial education purposes only. Not investment advice.