Rarely have I come across an investment thesis capable of mapping with such precision the transformative axis of an entire industrial era before it becomes legible to the majority. When this happens, the underlying logic is not the product of optimism but of an analysis capable of anticipating the real bottlenecks of a technological transition. Artificial intelligence is today the most discussed variable in financial markets and corporate strategic agendas, but the question most observers continue to underestimate concerns what makes AI scaling possible: not the models, not the algorithms, but the physical capacity to sustain them.
Leopold Aschenbrenner is 24 years old when he launches Situational Awareness LP, a hedge fund built around a precise thesis: the real bottleneck in the race toward artificial intelligence will not be software, but physical infrastructure — energy, chips, data centres, and large-scale computational capacity. The fund grows approximately 270% net of fees in the first months of 2025 and over 1,000% since its inception.
The story of Situational Awareness LP is inseparable from its founder’s. Aschenbrenner graduates as valedictorian from Columbia University at 19, joins OpenAI’s Superalignment team at 22 — working on the alignment of superintelligent AI systems — and is subsequently dismissed after circulating internal memos on safety and security risks. In June 2024, he publishes a 165-page manifesto predicting the emergence of AGI around 2027 as a direct result of extrapolating current trends, followed by a rapid transition toward superintelligence through AI research automation and recursive acceleration. On this thesis he constructs the fund’s portfolio: not AI applications, not large model companies, but the physical infrastructure that makes scaling possible — energy production including Bloom Energy, AI cloud computing and GPU hosting such as CoreWeave and Nebius, and Bitcoin miners converting operations to AI data centres including Core Scientific and IREN. The underlying logic is that the geopolitical competition between the US and China will accelerate infrastructure demand beyond what markets have already priced in, and that whoever controls the gigawatts and the electrical grid will control the next phase of the digital era. It is a position that transforms a hypothesis about how AI will develop into a capital allocation strategy with a precise and verifiable time horizon.
Menu
