A preview of how the AI-connected stock market bubble could burst and set off a broader financial crisis has been provided by the collapse of the once high-flying hedge fund Situational Awareness last week.

The demise of the fund was the outcome of the global loss in the market value of semiconductor stocks of $3 trillion in July.

Situational Awareness had made massive bets on the AI boom, leveraging its profits through large borrowings from the major banks, and was facing margin calls—demands for additional collateral—which would have led to a fire sale of its assets, possibly causing a major plunge across the market.

According to the Financial Times, had fire sales proceeded the “semiconductor rout” could have run “out of control.” It was only prevented when the hedge fund Citadel, owned by the billionaire Ken Griffin, moved in and bought $16 billion worth of Situational’s public equity holdings at a discount of 10 percent on Thursday. The market then breathed a sigh of relief and bounced up on Friday, with Citadel making a tidy profit on its deal.

But what took place could well be repeated on an even broader scale. After the rescue operation the New York Times commented that while AI-related stocks rallied and the market rose “that’s being treated on Wall Street as evidence of a reprieve rather than permanent proof the implosion will be contained.”

This is because, while it was an extreme case, the activities of Situational were not fundamentally different from what is taking place across Wall Street—the use of borrowed funds in order to juice profits to ever higher levels in the expectation that share prices will continue to rise. In the case of Situational, it has been estimated that for every dollar it provided from its own capital it borrowed an additional three or four dollars to make its trades.

The use of such leverage boosts the return on the upside but leads to major losses when the assets fall in value.

Leopold Aschenbrenner’s fundamental thesis was set out in an essay published in 2024 entitled “Situational Awareness,” the name he then used for his hedge fund established in that year. He maintained that AI demand would continue to grow requiring the massive building of data centres, the manufacture of chips and the development of electricity infrastructure—the same basic scenario that has powered the entire market.

Launching his fund in 2024, he wrote: “Before long, the world will wake up. But right now, there are perhaps a few hundred people, most of them in San Francisco and the AI labs that have situational awareness. Through whatever peculiar forces of fate, I have found myself amongst them.” And some of the biggest names on Wall Street bought into the claims.

The words were different, but the message was not essentially different from that with which Elon Musk launched the initial public offering (IPO) for SpaceX in June.

As stocks skyrocketed in the first half of the year, Situational was the toast of Wall Street and Aschenbrenner was being hailed as a visionary and Wunderkind as its market value climbed more than 400 percent. As recently as June he was described as the “Nostradamus of AI.”

These assessments were backed up by money from Goldman Sachs and others, including JPMorgan, Citigroup and Bank of America. It also received backing from the major financial trading firm Jane Street which rarely provides money to outside managers.

Everything was on the up and up and in early July Situational was selected to play a major role in $27 billion listing of the South Korean chipmaking firm SK Hynix on Wall Street.

But the hedge fund was then caught in the savage downturn in chip stocks in July, centred in South Korea, as doubts began to surface about the sustainability of the AI boom. These included when and even if the massive investment in data centres would turn a profit and over the problems for the American hyperscalers created by competition from China.

Situational reported a 67 percent loss in the value of the assets in the stocks that it held in July prompting margin calls from its financial backers. It held hundreds of millions of dollars in Bloom Energy and Sandisk which dropped by about 40 percent from the beginning of June. Stocks in Oracle and AMD in which it had significant holdings were each down by about 20 percent in July.

Overall, according to the NASDAQ index of global semiconductor stocks, there was a $3 trillion loss in market value in July.

Another indication of the decline is the fall in the market value of SpaceX. Since its IPO in mid-June is shares have halved in value from their peak and are now below their starting price.

There is a recognition in financial markets and media circles that the demise of Situational has a broad significance and confirms warnings by the Bank for International Settlements (BIS).

Together with other financial bodies, the BIS has noted that market valuations are overstretched and that the use of leverage to boost profits poses significant dangers in the event of a downturn. The fear is that margin calls by banks lead to fire sales of assets leading to a further downturn, with an impact on banks and other bodies which have provided the funds for speculation.

The use of leverage has been particularly pronounced in South Korea, where stock market largely rests on the chipmaker SK Hynix and Samsung.

Rapid falls in the Kospi index, transmitted to markets around the world because of their ever-closer integration, have been triggered by margin calls leading to the suspending of trading at least nine times this year “sending South Korea’s market into a state of chaos,” according to a CNN report.

 It said the “meteoric rise and near-collapse” of Situational Awareness had “sent shock waves across global markets, becoming a kind of cautionary tale for the AI era.”

Others are drawing the same conclusion. A Times article noted that Daniel Loeb, a well-known US investor and hedge fund manager had posted on X a reference to the collapse of the hedge fund Long Term Capital Management (LTCM) in September 1998.

That fund was bailed out to the tune of more than $3.6 billion—small change in comparison to the money which is coursing through markets today—in an operation organised by the New York Fed because of fears its demise could bring down the entire market.

In the event that did not happen. But the LTCM collapse was a warning of the developing crisis which was to erupt in September 2008. The collapse of Situational Awareness signals that another crisis, even bigger than 2008, could well in the making.