Leopold Aschenbrenner's Situational Awareness AI hedge fund disclosed a 49% allocation to memory semiconductor equities in its most recent 13F filing, positioning taken before the sector suffered a 67% peak-to-trough decline. The filing, covering holdings as of December 31, 2024, shows concentrated exposure to Micron Technology and SK Hynix—both central to the high-bandwidth memory thesis underpinning frontier AI training infrastructure.
The fund's US equity portfolio tilted heavily into HBM (high-bandwidth memory) manufacturers on the view that compute scaling for large language models would drive insatiable demand for specialized DRAM. Micron represented the largest single position. SK Hynix, the dominant supplier of HBM3E to Nvidia, accounted for the second-largest weight. The positioning reflected Aschenbrenner's public thesis that AI compute buildout would create supply constraints in advanced memory through 2025. That thesis held through year-end, then fractured in January when DeepSeek's R1 model demonstrated that inference efficiency could substitute for raw parameter scale, collapsing the memory demand curve allocators had priced in.
Memory stocks began their decline on January 27, hours after DeepSeek's technical report circulated among semiconductor analysts. Micron fell 17% in two sessions. SK Hynix dropped 21% over the same window. The sector-wide rout continued through February as every major Wall Street house revised HBM revenue estimates downward, citing reduced capex visibility from hyperscalers. The 67% drawdown figure—from sector highs in December to the February 14 trough—landed hardest on funds that had layered into memory on margin through Q4. Aschenbrenner's fund, launched in mid-2024 with a narrow mandate around AI infrastructure, had no hedges disclosed in the 13F and no offsetting short positions in low-margin commodity DRAM.
The disclosure matters because Aschenbrenner is not a portfolio manager by training. He is a former OpenAI researcher and author of the widely-read *Situational Awareness* essay on AGI timelines, which argued that whoever controls the compute stack controls the path to superintelligence. That essay became a fundraising document. The fund's LP base skews toward family offices and tech founders who bought the compute-scarcity narrative without stress-testing the memory pricing cycle. The 13F now functions as a case study in thematic conviction without sector hedging. Allocators who entered the fund in Q3 2024—when Aschenbrenner was still doing the conference circuit—are facing a 38% drawdown from their entry NAV, assuming they came in at average Q3 pricing and held through mid-February.
What makes the positioning unusual is the absence of diversification within the AI stack. No exposure to Nvidia. No positions in cloud hyperscalers. No offsets in power infrastructure or data center REITs. The 13F reads like a single-thesis bet: HBM as the choke point. That thesis was consensus in September. It became a crowded trade in November. It unwound in January. The fund's structure—hedge fund fees on a venture-style thesis—meant LPs paid 2-and-20 for exposure they could have accessed through sector ETFs at 0.4% annual expense.
Operators should watch for amended 13F-HR filings or voluntary disclosures in March that might show whether the fund added to positions during the January selloff or cut exposure. The next mandatory filing deadline is May 15, covering Q1 2025 holdings. If the fund held through the drawdown, it signals continued belief in the HBM scarcity thesis despite DeepSeek's efficiency gains. If it liquidated, that's a data point on how thematic funds behave when their core assumption breaks. Either way, the disclosed concentration gives allocators a clean before-and-after on what happens when a thought leader's essay becomes a portfolio.
The February trough in memory stocks has since recovered 14% as short-covering and dip-buying rotated in, but the sector remains 53% below December highs. Micron trades at 11x forward earnings, down from 18x in December. SK Hynix is at 9x, compressed from 15x. The multiple contraction reflects a market repricing AI memory demand as elastic rather than inelastic—a shift that invalidates the original Aschenbrenner positioning but may create entry points for allocators with longer horizons and tighter risk controls.
The takeaway
Thematic conviction without hedging: 49% memory concentration before 67% sector rout exposes LP cost of paying hedge-fund fees for venture-style bets.
Want the 60-second program for your specific event?
Enter your event and email — we build it and send the branded proposal before lunch. No obligation.
The branded-identity layer Chiefs of Staff and heritage CMOs route through — your name imprinted on real authorized stock, your pick of 200+ brands and 70,000 products, shipped from one accountable house. Nine editorial desks publish the intelligence those operators read before they sign.
200+authorized brands
70,000products · virtual proof on each
9 deskspublishing daily
1997one house, since
70,000 SKUs · virtual proof in 60 seconds · no platform fee · blind-shipped · ASI #217876
Your next customer won't visit your website. Their AI will.
AI assistants have quietly taken over the first step of buying — they answer from catalogs they can read and shortlist whoever can actually ship. Two questions now decide whether you exist to that buyer: can a machine read your catalog, and can you fulfill the order. Most brands fail one or both and never find out why the orders went elsewhere. The winners of this shift aren't the loudest. They're the most readable. Build for the machine that's about to do the shopping.
Built by the craft floor — apparel, media, packaging, and secure print.
This trade runs on hands, not desks. Imprint manufacturing & Komori Press · Canon high-speed secure-media operations is a craft floor — genuine Six Sigma discipline applied to ink, thread, foil, and registration, where a hundredth of an inch is the difference between a brand that reads serious and one that reads cheap. POPS4 is built by exactly those operators: independent, boots-on-the-ground engineers who carry their own book, read a client in microseconds, and put their name on every run. Beyond our own Virginia Beach floor, we work with a vetted network of craft manufacturers across the US — each meeting the highest excellence in QC standards in the industry, each a specialist in its own discipline — so apparel, hard-goods imprinting, media manufacturing, packaging, and secure printing all go to the bench built for them, coordinated from one accountable hub. Short-run from twenty-five units, volume to five hundred thousand. Two hundred authorized national brands, seventy thousand SKUs with virtual proofing on every one. Art archived for instant reorders. Net-thirty corporate terms, NDA-standard white-label — your name on the work, or none at all.
Strategy, positioning, identity, creative, and messaging — wired into an AI system that publishes and distributes on its own. Nine editorial desks generate the authority, the production house ships the physical proof, and the attribution layer tells you which post sold which SKU. What you get is an operating layer — content, catalog, and order path under one roof — that keeps working whether or not you are in the room. Built for principals who would rather own the machine than rent the agency.
Named-account programs — one desk, quiet delivery, NDA-standard.
One point of contact who already knows the file, so nothing restarts from zero between engagements. The work ships blind, under NDA, with your name on it or none at all. Built for single-family offices, heritage-house CMOs, sports-ownership groups, and the agencies that white-label our production. The relationship is the product; the merch is the proof of it.
SFO · Chief of Staff desk. Principal household, properties, aircraft, yacht, calendar, philanthropy — one file.
Shop seventy thousand products. Virtual proof on every one. 24/7.
Drop your logo on any product and see the virtual proof before asking. Quote routes direct to the desk. MCP catalog for AI agents. Celeste for the fast conversation. Full self-service checkout in development.