Situational Awareness collapsed from $45 billion to $10 billion in a matter of days—a whopping 77% wipeout—thanks to highly leveraged bets on volatile AI infrastructure stocks like Nebius, SanDisk and CoreWeave.

Getty

In June 2024, former OpenAI researcher Leopold Aschenbrenner went viral for a behemoth, 165-page essay titled “Situational Awareness: The Decade Ahead,” which predicted AI would become more powerful than humans. The 25-year-old decided to convert that foresightedness into fortune. Within months, he set up a hedge fund named for the essay to invest in public companies benefiting from the AI boom and prominent AI startups. Most investors were required to put in a minimum of $25 million into the fund, which drew backing from elite financial institutions like JPMorgan, Goldman Sachs and Bank of America, as well as billionaires like Stripe cofounders Patrick and John Collison, the Wall Street Journal reported. The fund had turned hundreds of millions of dollars into $20 billion in assets by mid-2026.

But by the end of July, investors got spooked by the amount of AI spending and markets had started to turn. The fund collapsed from $45 billion to $10 billion in a matter of days—a whopping 77% wipeout—thanks to highly leveraged bets on volatile AI infrastructure stocks like Nebius, SanDisk and CoreWeave. Billionaire Ken Griffin’s firm Citadel stepped in to bail out Aschenbrenner, buying up the bulk of its positions at a steep 10% discount, according to the WSJ report. By the end of the fire sale, Situational Awareness was left with a slimmed down portfolio of stocks and startups valued over $10 billion, including a significant stake in Anthropic. The sudden vaporization of the young AI investor’s high-flying portfolio points to how a lack of risk management can burn a lot of money, very fast.

I’m always looking for ways to improve this newsletter. What is one thing you’d love to see more or less of in The Prompt? I’d love to hear from you at rshrivastava@forbes.com.

Now let’s get into the headlines.

ETHICS + LAW

Apple’s lawsuit against OpenAI just got messier. In early July, the iPhone maker sued OpenAI and two of its former employees, accusing them of stealing trade secrets and retrieving confidential information after leaving. In a lengthy response to the lawsuit, OpenAI rejected Apple’s claims, noting that its lawyers reached out to the wrong person after confusing two Asian last names, and one of the accused former employees accessed information only after Apple employees asked for his help in locating it. “Apple is one of the greatest companies of all time, and built a reputation for obsessing over the smallest details. This careless, aggressive and oddly personal lawsuit sadly doesn’t live up to that reputation,” the company wrote in a blog.

PEAK PERFORMANCE

Chinese tech behemoth Alibaba announced a new AI model Monday that it claims is on par with Anthropic’s Fable and OpenAI’s latest GPT models. The model, scheduled for release next week, is built for coding and other types of knowledge work. In an internal test, the model spent 16 days building and fixing up an AI coding tool.

AI DEAL OF THE WEEK

Flapping Airplanes, an AI startup that’s researching new ways to train AI models using less data, is in talks to raise hundreds of millions of dollars at a $5 billion valuation, sources told Forbes last week. The company is among a growing number of so-called neolabs that have raised massive sums of money to carry out research instead of building and commercializing products.

DEEP DIVE: Silicon Valley’s Other China Problem: It’s Training Their AI

When U.S. sales teams from Silicon Valley’s big data-labeling startups visited this year’s International Conference on Machine Learning in Seoul, Korea, they arrived prepped to court the industry’s big spenders. The data companies—collectively worth tens of billions of dollars and generating billions in annual revenue supplying training data to customers like OpenAI and Anthropic—are used to chasing AI labs that are notoriously demanding, fickle and difficult to satisfy.

They found another eager customer waiting for them: China’s AI industry.

Some Chinese companies have shopping lists. Tencent—which has previously been designated by the U.S. government as associated with the Chinese military, a characterization the company disputes— circulated with prospective vendors a detailed request for training data spanning finance, cybersecurity and one of AI’s most coveted research goals: AI systems capable of improving themselves.

While the U.S. has long prohibited China from buying the chips that power America’s top-tier AI models, it hasn’t taken the same precautions with the data required to train them. This packaged human expertise, produced by specialist networks and shaped by task designs, grading rubrics and quality controls, is part of the expensive data infrastructure that teaches models to handle difficult professional tasks like financial modeling and coding. The business is worth hundreds of millions of dollars annually and it’s helping Chinese AI models close the gap with their American rivals.

Documentation and communications from buyers at Chinese AI labs and employees from top Silicon Valley data platforms such as Surge AI and Mercor reviewed by Forbes reveal a quiet multi-hundred-million-dollar trade in training datasets. The same U.S. startups serving as suppliers to OpenAI, Anthropic, and, in some cases, U.S. federal entities, are also supplying Beijing’s top labs. Increasingly, Silicon Valley is training both sides of the AI arms race.

Read the full story on Forbes.

MODEL BEHAVIOR

Job interviews typically require candidates to sit for multiple rounds of stress tests and personality assessments. But early AI startup LemonLime took its recruiting to the extreme: it offered attendees at a networking party the chance to interview if they got a permanent tattoo of the company logo. Seven people took the offer. After the move sparked online backlash, founder and CEO Jordan Zietz publicly apologized and said that the idea was to “meet exceptional people, and find out which of them are just as crazy as we are.”