AI-focused fund linked to former OpenAI researcher reportedly trims public equity bets, with NVDA, AMD, ORCL and MU among positions in focus

Markets news posts circulating Tuesday said a $20-billion AI-focused fund associated with former OpenAI researcher Leopold Aschenbrenner is moving to reduce its exposure to public equities. The reporting characterizes the move as an effort to unwind a large portion of the fund’s listed-stock holdings as the AI trade faces renewed pressure after a broader sector selloff.

The same market chatter said the fund is selling a “bulk” of its public portfolio to Citadel. The posts did not lay out the timing, the exact percentage of holdings being cut, or whether the sales are part of a larger liquidity or rebalancing plan, leaving many specifics unclear.

Among the individual stocks cited as being in focus were Nvidia, AMD, Oracle, and Micron Technology. All four are widely held names with substantial ties to data-center buildouts and the supply chain feeding AI and cloud computing, which can make them sensitive to changes in investor sentiment and expectations for technology spending.

For semiconductor and related infrastructure investors, the reported de-risking matters because it suggests at least one large, AI-oriented allocation is recalibrating while valuations and trading momentum remain under strain. Even without detailed order data, a reduction at the portfolio level can translate into incremental selling pressure across highly liquid large-cap equities.

The reported scale of the fund, described as $20 billion, also highlights how portfolio concentration in AI beneficiaries can become a focal point when market conditions shift. When large funds rebalance during drawdowns, the market often looks for indicates in the most crowded positions, particularly in semiconductors and enterprise infrastructure stocks that investors treat as proxies for AI demand.

Beyond the individual tickers, the episode fits a broader pattern in markets where AI-linked exposure has moved quickly as macro and industry assumptions change. In recent months, investors have grappled with questions about the pace of AI infrastructure spending, competition, and the cost of building and powering data-center systems, all of which can amplify volatility in “picks-and-shovels” and chip-related shares.

Still, investors may want to wait before treating the report as a definitive read-through for near-term corporate performance. The market post did not disclose whether the underlying companies’ fundamentals changed, whether the sales were motivated by valuation, risk limits, or a shift in strategy, or how much of the portfolio remains invested after the alleged trims.

For now, the most watchful announcement will be follow-through: whether additional disclosures, filings, or subsequent portfolio updates corroborate the reported unwind. In the absence of such detail, the key takeaway is the same across AI-heavy names: public-market positioning can change quickly when sentiment turns, and the stocks most associated with the trade are often the first to show up in “in focus” lists during selloffs.