Leopold Aschenbrenner
FOLLOWSituational Awareness LP
The most data-driven AGI bull in the world now runs a $5.5B hedge fund -- his 13F portfolio is a roadmap of AGI infrastructure plays (data centers, power, chips, networking) worth tracking for our thesis.
AI/Tech-Aware Value Investors & AGI Thinkers
Score Breakdown
Investment Philosophy & Portfolio Style
Philosophy
Aschenbrenner's investment thesis is entirely derived from his AGI timeline analysis. His core prediction: AGI will arrive by 2027, with superintelligence following within 1-2 years. His 'Situational Awareness' essay argues that the path from GPT-2 to GPT-4 (preschooler to smart high-schooler) took 4 years, and comparable progress will continue through 2027 driven by compute scaling (~0.5 orders of magnitude annually), algorithmic improvements, and 'unhobbling' gains. Once AGI is achieved, he predicts hundreds of millions of AI agents could automate AI research itself, compressing a decade of algorithmic progress into less than a year. His investment thesis centers on: (1) unprecedented demand for compute infrastructure -- 'trillions of dollars' will flow into data centers, GPUs, and power infrastructure; (2) American electricity production must grow by 'tens of percent' by decade's end; (3) data center operators, power providers, networking equipment, and compute infrastructure companies are the bottleneck and therefore the most investable theme. His actual portfolio confirms this: the entire $5.5B is concentrated in AI infrastructure -- data centers (CoreWeave, Core Scientific, IREN, Applied Digital, Cipher Mining), energy/power (Bloom Energy, EQT Corp, Solaris Energy), semiconductors/chips (Intel calls, Lumentum, Coherent, Tower Semiconductor, SanDisk), and Bitcoin mining/compute operators pivoting to AI (Riot, CleanSpark, Bitfarms, Hut 8, Bitdeer). This is thesis-driven thematic investing, not traditional value investing.
Portfolio Style
Highly concentrated thematic portfolio with 25 positions totaling $5.5B as of Q4 2025. The top 7 positions make up 79% of the portfolio. Heavy use of call options alongside equity for several positions (CoreWeave, Intel, Bloom Energy, EQT), amplifying upside exposure. Portfolio themes: (1) AI data center operators: CoreWeave 22%, Core Scientific 7.6%, IREN 6%, Applied Digital 5%, Cipher Mining 2.8%; (2) Power/energy: Bloom Energy 16.5%, EQT 3.1%, Solaris Energy 1.6%; (3) Networking/semiconductors: Intel 13.5% (almost entirely call options), Lumentum 8.7%, Coherent 1.6%, Tower Semi 1.5%, SanDisk 4.5%; (4) Bitcoin miners pivoting to AI compute: Riot 1.4%, CleanSpark 0.3%, Bitfarms 0.3%, Hut 8 0.7%, Bitdeer 0.4%; (5) Real estate: Kilroy Realty 0.9% (San Francisco office REIT). Notable: the Intel position ($747M) is almost entirely call options ($747M calls vs $37 in equity -- literally 1 share), suggesting a speculative bet on Intel's turnaround as an AGI chip beneficiary. The Infosys put ($9M) appears to be a short/hedge on IT services companies that would be disrupted by AGI. The entire portfolio is a single macro bet: AGI is coming, and the bottleneck is infrastructure.
Background
Leopold Aschenbrenner (born ~2000) is a former OpenAI researcher who became one of the most influential voices on AGI timelines. He studied economics at Columbia University, graduating at age 19, and was a research fellow at the Global Priorities Institute at Oxford before joining OpenAI's Superalignment team under Ilya Sutskever. In mid-2024, he was fired from OpenAI after raising internal concerns about security practices. Shortly after, he published 'Situational Awareness: The Decade Ahead,' a ~165-page essay that became the most widely-read analysis of AGI timelines in 2024. Following the essay's success, Aschenbrenner founded Situational Awareness LP, a Delaware-based hedge fund headquartered at 512 2nd Street, Suite 400, San Francisco, CA 94107. The fund raised over $1 billion from 66 accredited investors ($5M minimum investment) with first sales in November 2024. As of Q4 2025, the fund manages approximately $5.5 billion in public equities per its 13F filing. The fund's general partner is SAF AI GP LP, with Carl Shulman (noted AI safety researcher) serving as co-portfolio manager, Nicholas Gross-Whitaker as COO, and Christopher Zellner as Chief Compliance Officer. Aschenbrenner serves as Managing Partner and control person. The fund filed its first 13F in February 2025 and has filed Schedule 13D for Core Scientific (CORZ) where it holds a 9.4% ownership stake.
Track Record
Very limited -- the fund began investing in November 2024, making it roughly 1.5 years old. No public performance data. The intellectual track record is more notable: 'Situational Awareness' has been remarkably prescient about the AI investment surge, the national security discourse around AI, and the scaling laws continuing to hold. However, his most aggressive prediction -- AGI by 2025-2027 -- remains unvalidated. The fund has grown rapidly from ~$1B in investor commitments to $5.5B in managed equities, suggesting strong early performance or additional capital raises. His 13D filing on Core Scientific reveals he spent $362M acquiring a 9.4% stake, indicating substantial conviction and willingness to take activist-style positions.
Notable Holdings
As of Q4 2025 13F (filed Feb 2026): CoreWeave (CRWV) 22% ($1.2B, shares + call options); Bloom Energy (BE) 16.5% ($911M, 10.1M shares + calls); Intel (INTC) 13.5% ($747M, almost entirely call options on 20.2M shares); Lumentum (LITE) 8.7% ($479M); Core Scientific (CORZ) 7.6% ($419M, 28.8M shares, 9.4% ownership per 13D); IREN (IREN) 6.0% ($329M); Applied Digital (APLD) 5.0% ($278M); SanDisk (SNDK) 4.5% ($250M). Also holds EQT, Cipher Mining, Coherent, Solaris Energy, Tower Semi, Kilroy Realty, and several Bitcoin miners. Put on Infosys (short IT services). Total: $5.5B across 25 positions.
Transparency & Integrity
Transparency(Score: 8/10)
High on intellectual framework, moderate on investment specifics. The 'Situational Awareness' essay is freely available and extraordinarily detailed. The fund now files SEC 13F quarterly (first filed February 2025), providing full visibility into US equity holdings. Schedule 13D filed on Core Scientific provides additional detail on thesis and ownership. Form D filings reveal fund structure, team, and total capital raised. However, investor letters and performance data are not public. He is active on social media (X/Twitter) and gives occasional interviews and podcast appearances (Dwarkesh Patel, etc.).
Integrity(Score: 8/10)
High. His whistleblowing at OpenAI -- raising concerns about security practices at personal cost (he was fired) -- demonstrates intellectual courage. His essay is intellectually honest, explicitly stating assumptions and allowing readers to disagree. The 13D filing on Core Scientific states the position was acquired 'based on their belief that the Common Stock, when purchased, was undervalued and represented an attractive investment opportunity' -- straightforward language. Carl Shulman's involvement as co-portfolio manager adds credibility (Shulman is a respected AI safety researcher, not a finance promoter). Potential conflict of interest noted: he published a widely-read essay predicting massive AI infrastructure buildout, then raised a fund to invest in exactly that. On balance, the depth and rigor of his analysis suggests genuine conviction.
Relevance to Us
Leopold Aschenbrenner is now highly relevant as both an intellectual guide AND an investment signal. His core thesis -- AGI by ~2027, with massive infrastructure investment required -- aligns closely with our own AGI-by-2030 assumption. His fund's actual 13F holdings now provide concrete, actionable intelligence about which specific companies he believes will benefit most from the AGI buildout. His portfolio is essentially a roadmap of AGI infrastructure plays. While his approach differs from ours (thesis-driven momentum rather than downside-first value), several of his positions are worth deep analysis: (1) Core Scientific (CORZ) -- he owns 9.4% of the company and spent $362M acquiring the stake; (2) Bloom Energy (BE) -- $911M bet on fuel cell power for data centers; (3) Lumentum (LITE) -- optical networking for data center interconnects; (4) Intel (INTC) calls -- a contrarian bet on Intel's AI chip revival. These positions give us specific companies to analyze for our own floor-price framework. If AGI is truly coming by 2027-2030, the infrastructure companies Aschenbrenner is betting on deserve our attention, even if we apply very different valuation methodology.