Ken Griffin, founder and CEO of multinational hedge fund Citadel.
Ken Griffin, founder and CEO of multinational hedge fund Citadel, speaks during the 29th annual Milken Institute Global Conference at the Beverly Hilton in Beverly Hills, California on May 5, 2026. Patrick T. Fallon/AFP

Citadel's flagship Wellington fund posted 5.9% returns in July, its largest in years, after the company bought assets from collapsed Situational Awareness at discounted prices.

CNBC detailed that the return is the best since 2022 and pushed this year's gains to 12%. It follows the decision of buying most stocks held by the collapsed funds after it saw large losses resulting from margin calls and forced asset sales. Citadel managed some $71 billion in assets on July 1.

Read more

Aschenbrenner's fund had bet that the AI boom would drive sustained demand for chips, memory, data centers and electricity.

The hedge fund had reportedly grown to approximately $45 billion at the beginning of July before the market downturn eroded the value of its portfolio. Situational Awareness had been reportedly negotiating to sell its stake in AI startup Anthropic.

The fund's largest publicly disclosed holdings at the end of the first quarter included Nebius Group, SanDisk, Micron and CoreWeave. All four companies had lost more than 35% of their market value before the latest rebound.

Aschenbrenner rose to prominence in 2024 after publishing a series of widely read essays arguing that increasingly capable AI systems would require an unprecedented expansion of computing infrastructure.

His writings emphasized growing demand for advanced semiconductors, high-bandwidth memory, power generation and data centers, themes that became central to both AI investing and Situational Awareness' portfolio construction.

The 25-year-old investor followed an unconventional path into finance. After graduating as valedictorian from Columbia University at just 19 years old, he joined OpenAI's Superalignment team, which focused on long-term AI safety.

His tenure at OpenAI ended in 2024 after the company dismissed him over what it described as an improper disclosure of internal information. Aschenbrenner has disputed that characterization, saying he shared a largely nonconfidential planning document with outside researchers for feedback.

He has also argued that his dismissal followed disagreements over security concerns he raised about OpenAI, while the company has maintained those concerns were unrelated to his departure.