Leopold Aschenbrenner's $2 Billion Bet on AI-Era Semiconductors: Why This Hedge Fund Is Shorting Industry-Wide Exposure

Former OpenAI researcher Leopold Aschenbrenner founded Situational Awareness, a hedge fund that has grown to over $2 billion in assets under management, specifically to identify market winners and losers shaped by artificial intelligence advancement. His investment strategy reveals a sophisticated approach to semiconductor sector exposure: simultaneously building protective hedges against broad industry decline while establishing concentrated long positions in two select chipmakers he believes will dominate the AI-driven future.

The fund’s most recent SEC 13F filing exposed a striking paradox in its portfolio construction. Leopold Aschenbrenner’s team holds massive put option positions representing a substantial short bet against the semiconductor industry as a whole, yet simultaneously maintains significant bullish positions in Intel and Broadcom. This dual approach reflects a nuanced understanding of how AI will reshape the chip landscape—not uniformly, but with clear winners and substantial losers emerging over the coming years.

Defensive Strategy: The Semiconductor ETF Hedge

At the end of the second quarter of 2025, Situational Awareness held 20,441 put contracts on the VanEck Semiconductor ETF (NASDAQ: SMH), the fund’s largest position by notional value. Put contracts grant the holder the right to sell shares at a predetermined strike price, increasing in value as the underlying asset declines. This represents 27% of the fund’s total public holdings.

The bet against the broader semiconductor industry makes strategic sense when considering that the ETF is cap-weighted, meaning that major players like Nvidia—which comprises over 20% of the ETF’s portfolio—dominate the index. Leopold Aschenbrenner’s thesis appears to be that while AI will create immense opportunity, most established chipmakers will face disruption from specialized competitors or fail to adapt their business models. The protective put position likely hedges against systematic risks to the overall sector that could negatively impact the fund’s core holdings.

However, this defensive posture tells only half the story. Leopold Aschenbrenner’s strategy isn’t bearish on all semiconductor companies; rather, it reflects conviction in specific narratives driving selective long positions accounting for 37% of the fund’s publicly traded portfolio.

Strategic Long Positions: Intel and Broadcom Lead the Way

Situational Awareness established a significant position in Intel call options during early 2025, followed by substantial stock holdings in Broadcom throughout the year. Call options, which increase in value if the underlying security rises, provide leveraged exposure to the fund manager’s thesis regarding these companies’ AI-era prospects.

The Intel Thesis: Geopolitical and Strategic Imperative

Leopold Aschenbrenner has written extensively about AI superintelligence development as a matter of national security. This framework shapes his confidence in Intel despite the chip designer’s recent struggles to maintain relevance in AI-optimized data centers. While GPUs and specialized accelerators have largely displaced Intel’s traditional CPUs in AI workloads, Aschenbrenner sees a critical role for Intel as the only leading-edge semiconductor foundry based in the United States.

His conviction has begun to materialize. The U.S. government agreed to acquire a 9.9% stake in Intel in late 2025, with additional warrants to purchase an additional 5% stake contingent on the company securing major manufacturing contracts. This government investment accompanied a major commitment from Intel to deploy $100 billion into U.S.-based semiconductor facilities, directly supporting foundry operations crucial for domestic AI chip production.

The Broadcom Thesis: Custom Silicon and Hyperscaler Partnerships

Leopold Aschenbrenner’s Broadcom position reflects a different investment narrative: the expectation that AI chips will become increasingly specialized and customized for specific workloads. Rather than relying on standard, off-the-shelf processors, hyperscale data center operators will increasingly deploy custom-designed silicon optimized for their particular AI infrastructure needs.

Broadcom has positioned itself as a critical partner in this shift, supporting four major hyperscale customers—including a recently announced partnership reportedly with OpenAI—representing $10 billion in contract commitments for 2026 alone. Management guidance suggests AI chip revenue will accelerate through 2026 as the company expands market share among both existing and newly added customers.

The Investment Thesis Behind Each Bet

The beauty of Leopold Aschenbrenner’s strategy lies in its internal consistency. Both long positions rest on fundamental shifts in how AI infrastructure develops: Intel as the geopolitically necessary domestic foundry provider, and Broadcom as the critical enabler of custom silicon proliferation. Meanwhile, the broad semiconductor hedge protects against companies caught in the transition—those unable to specialize or adapt their business models to an AI-centric infrastructure paradigm.

It’s important to note that this analysis reflects the fund’s positioning through mid-2025. By early 2026, both Intel and Broadcom stock prices have appreciated significantly on developments supporting Leopold Aschenbrenner’s theses, meaning new investors face substantially higher entry valuations. While the long-term investment narratives remain intact, the near-term risk-reward calculus has shifted considerably.

What This Means for Semiconductor Investors

Leopold Aschenbrenner’s hedge fund construction offers a masterclass in navigating sector transitions driven by transformative technologies. The lesson isn’t necessarily that investors should replicate every position, but rather that the semiconductor sector won’t produce uniform winners across all players. AI will accelerate the separation between companies that successfully adapt and those that face margin compression or market share loss.

The $2 billion fund’s strategic positioning suggests that investors concerned about broad semiconductor exposure might consider how their portfolio construction aligns with these emerging structural trends. Leopold Aschenbrenner’s willingness to simultaneously short the sector while establishing concentrated long positions in select names reflects confidence that this divergence will only intensify as AI infrastructure development accelerates through 2026 and beyond.