Situational Awareness LP, run by former OpenAI researcher Leopold Aschenbrenner, disclosed roughly $5.52 billion in U.S. equity exposure across 29 positions in its Q4 2025 13F filing, with holdings stated as of September 30, 2025. The filing reads like a concentrated bet on the intersection of Bitcoin mining and AI infrastructure, rather than a broad-market portfolio.
Alongside the 13F, the disclosures included an amended Schedule 13D that recorded a near 9.4% stake in Core Scientific, described as 28,756,478 shares with shared voting and disposition power. That combination of scale and governance footprint raises the temperature for counterparties, because it signals not only exposure, but also potential influence.
Portfolio posture and the AI-infrastructure angle
The portfolio’s largest reported exposures underscore that positioning. CoreWeave was listed around $563 million, while Core Scientific was reported around $362 million, and Iris Energy around $338 million, based on the same September 30, 2025 snapshot. Even at a headline level, the allocations cluster around power-dense, data-center-adjacent businesses that can be framed as compute infrastructure.
The filing also showed Situational Awareness holding positions in Cipher Mining, Hut 8, Riot Platforms, Bitfarms, Galaxy Digital, and Bitdeer, while reducing exposure to Applied Digital. The overall footprint reinforces a thesis-driven portfolio construction, where miners and adjacent operators are treated as the picks-and-shovels layer for compute demand.
Aschenbrenner’s stated rationale ties the package together: “the true beneficiaries of the artificial intelligence revolution will not be the AI developers themselves, but rather the foundational infrastructure that powers it.” In that framing, mining sites and data-center assets become convertible capacity—the kind of platform asset that can be redeployed toward high-performance compute and AI hosting.
What changes for compliance, governance, and counterparties
For compliance teams and institutional counterparties, the key shift is the implied reclassification of miners from crypto operators to energy-and-compute infrastructure plays. That reframing can force updates to counterparty assessments, especially around power procurement, colocation dependencies, and the durability of revenue narratives used in diligence and reporting.
The Core Scientific stake is the clearest governance signal. A 13D-sized position can introduce heightened disclosure sensitivity and governance oversight expectations, and it often prompts tighter monitoring of subsequent amendments, voting behavior, and any changes in stated intent. When an investor holds both material economic exposure and formal influence levers, the market typically demands more transparency and cleaner lines of accountability.
Operationally, the thesis implies real balance-sheet and execution consequences for the underlying companies. A miner that pivots toward AI hosting may need to renegotiate energy contracts, change capex priorities, and rebalance asset composition, which can affect financial reporting quality and the risk profile that counterparties underwrite.
For custodians and asset managers, the convergence of crypto mining and AI compute complicates classification and control frameworks, especially when revenue streams and operational risks diverge inside the same corporate wrapper. Segregation, audit-grade record-keeping, and due-diligence artifacts need to keep pace, or counterparties risk underwriting the wrong business model.
Finally, it’s important to treat these disclosures as a time-stamped snapshot. The holdings were reported as of September 30, 2025, and positioning may have changed since the filings were submitted. The practical next step for market participants is straightforward: monitor subsequent SEC filings, any further Schedule 13D amendments, and company disclosures to see whether the AI-infrastructure pivot thesis translates into sustained operational focus and whether voting power or exposure materially shifts.