Scope · AFS
Of Leopold Aschenbrenner / Situational Awareness LP, for you. Cites the locked monthly. Does not replace §00. Not a buy list.
Evidence base
Scout corpus aschenbrenner-situational-awareness-2026-07. SoT: July 31 LP letter + on-record wires (CNBC, Reuters). There is no public live portfolio. Delayed 13F is used only for structure (position counts, notional, puts on then off) — not as the July book, and not as a holdings table.
August snapshot
Your 11: /2026-08. Pair case study: Druckenmiller. Cycle engine: BTC bear. Coefficient stays 0.
Verdict — theme partially survived; expression and structure did not
What happened: a July month in which AI-infra longs fell, a software short (Adobe, named on the record) rallied, reported leverage up to 4× (press, not a filing) turned it into a near-impairment, and the public book went to Citadel in a block. What went right: the fund’s own letter still holds the infra-scarcity theme as direction over months. What went wrong: expression and structure together — not “all right, just over-levered.”
Pair: Druckenmiller reversed vehicles without a prime-broker gun. Your 11 live on the August monthly.
What happened
Situational Awareness LP, inception through the July 31 letter. The live book is not public. Spine: March short overlay → June larger long, overlay off → July spiral. Structure counts are not a holdings list.
| Window | What we can say | Source |
|---|---|---|
| 2024 | Essay Situational Awareness: The Decade Ahead — the theme, not the 2026 fund’s structure. | Essay [A] |
| Q4 2025 as-of 31 Dec | Visible long sleeve: $5.52B, 29 positions, 100% long. Structure photograph, not the live book. | 13F structure [A] |
| March · Q1 as-of 31 Mar | Short overlay on: $13.68B, 42 positions. Puts ~62% of notional across eleven names. All eleven rows are gone by the next filing. | 13F structure [A] |
| June · Q2 as-of 30 Jun | Overlay off. $20.24B, 26 positions, 100% long. Last public snapshot — one month before July. Book notional ~4× Q4 in six months. | 13F structure [A] |
| Week of 27 Jul | Both legs hurt: AI-infra longs down; Adobe named as a short that rallied against him. The July short is not in any 13F Scout can source. | CNBC [A] |
| ~29–30 Jul | Public book sold in a block to Citadel. Primes named: BofA, GS, JPM, Citi. | CNBC / Reuters [A] |
| ~29–30 Jul | Anthropic-stake sale agreed then withdrawn. | WSJ pickup [B] |
| 31 Jul letter | −67% July MTD · +80% YTD, both unaudited estimates. Shorts closed. Portfolio financing removed. Fully-paid-for public book. Fund not shut. Hybrid public-private. Anthropic stake kept. | LP letter [A] |
Treat “forced fire sale” vs the letter’s “decisive action” as unresolved. Prefer the letter when precision matters. No ticker table on this page.
Why
The fund’s name is the manifesto. Theme-as-identity; leverage as amplifier; the July expression was not a clean directional tape.
- Theme-as-identity — little daylight between “AGI scarcity is right” and “this fund should be large.”
- A hedge/short overlay is visible as structure in Q1 (puts on, then off by Q2). The July short that actually hurt (software) is not in any public holdings file.
- Leverage (press: up to 400%) stacked on concentration. The letter’s fix is a fully-paid-for public book.
- AUM and fame scaled faster than disclosed risk governance. Operating maturity is the open question, not a ticker list.
Right / wrong
| Dimension | Score |
|---|---|
| Theme | Partially right. Letter: fundamentals accelerating as prices fell. A two-week dislocation does not kill infra scarcity. |
| Expression | Wrong. Both legs lost together (CNBC). The July software short has no public size or entry. Crowded, correlated, not a clean “AI trade.” |
| Structure | Catastrophic as the converter of a bad month into an existential month. Letter: a drawdown should never jeopardize the fund. |
Composite
All three contributed. Dropping either expression or leverage erases how a survivable miss became a block sale (the deal itself is [A]; a ~24-hour window is press reconstruction [B]). Pair vs Druckenmiller: regime-before-vehicle vs thesis-as-identity; concentration-as-attention vs concentration-plus-undisclosed-leverage; voluntary full-exit vs exit under financing pressure.
Against our data
No Cloud QUANT blotter row is this fund. There is no public live book to match against. Do not fade a name because a recap site printed a delayed filing.
Process against for QUANT: high conviction is not a risk budget; a hedge leg needs an exit plan; theme ≠ this trade. No falsifier stub until a live P-* actually speaks. Our tape first.
Dataset gaps
| Gap | Ours? |
|---|---|
| G1 | The July book itself — longs, shorts, sizes, at the margin call. Never-ours. No public live portfolio exists. |
| G2 | Leverage ratio, margin terms, which prime held what. Press 400% is not a filing. Never-ours (Form PF is confidential). |
| G3 | Adobe (and any other) short size/timing. Named in wires only. |
| G4 | One named stake via SC 13G (Nebius) — an ownership notice, not a portfolio. Scout has not re-fetched the 13G text this round. |
| G5 | Forced vs the letter’s “decisive action.” Prefer the letter. |
| G6 | Audited July/YTD. Letter says unaudited estimates. |
Do not hunt a 13F for a buy list. Private Anthropic marks stay out of scope. The only other holdings-class filing in the Scout file is a possible SC 13G on one name — still not the book.
Appendix — no public live book
Unlike Duquesne, the July trade sat in shorts, leverage, and intra-month fills. Those never land in a 13F. Structure counts below are the delayed long sleeve, not the July book.
- LP letters — P&L, leverage policy, “fully-paid-for,” fund still open. Not a name list.
- Wires — Citadel block, primes, Adobe named as a short, both legs hurt.
- 13F structure [A] — 29 → 42 → 26 positions; $5.52B → $13.68B → $20.24B; puts on then off. Not a holdings table.
- SC 13G — can flag a single large long. Not a portfolio.
- Form PF / ADV / primes — PF is confidential. ADV is not holdings. Prime books are private. Citadel now holds the old public sleeve.
Scout holds EDGAR 13F XML in the corpus. Names stay in the corpus so this page cannot be misread as his book.
Notes
- Case study — of Situational Awareness / Aschenbrenner, for you. Letters, not a 13F book.
- Sources — Scout corpus
aschenbrenner-situational-awareness-2026-07· July 31 letter quotes via Business Insider/AOL match · CNBC / Reuters on-record · 13F for structure counts only.