The power to see the future, and learn from the past

Choy Jung-hyuck

The author is a professor of economics, finance and asset management at Hanyang Cyber University.

A hedge fund named Situational Awareness ironically failed to grasp its own situation. Founded by former OpenAI researcher Leopold Aschenbrenner, the fund gained 270 percent through May by concentrating on AI-related stocks. But insight proved no match for excessive leverage. When AI shares tumbled, including major holding SK hynix, margin calls forced the fund to sell stocks for cash. Its value fell 67 percent in July.

Intriguingly, most of its publicly traded stock portfolio was acquired by Citadel, the hedge fund led by Ken Griffin. Griffin built Citadel into a global firm managing more than $70 billion. Yet he, too, faced an existential challenge during the 2008 financial crisis, when heavy losses created severe liquidity strains. He later strengthened risk controls, particularly over leverage and liquidity, helping turn Citadel into a firm capable of exploiting crises rather than merely surviving them. That experience became a lasting reminder that liquidity can matter as much as being right.

Griffin's investment philosophy was evident in a Goldman Sachs interview in June. Investors cannot prepare for every extreme risk, he said, so they must manage positions by asking whether they can survive the worst-case loss and remain able to fight back. The point is not to predict everything markets might do but to remain in the market when predictions fail. Weeks later, Citadel bought Situational Awareness's forced-sale portfolio at a reported 10 percent discount from market prices.

Strong conviction can easily lead to concentrated positions and greater leverage. Leverage rapidly magnifies gains in rising markets but strips investors of choices when markets move against them. Its greatest danger is not merely larger losses. Even when a long-term thesis is correct, leverage can prevent investors from waiting for it to materialize. Aschenbrenner may have had exceptional insight into AI's future, but he could not predict the path stock prices would take to get there.

Believing AI will transform the world over the long term is not the same as believing AI-related stocks will keep rising. Long-term investing is less about avoiding corrections than building a structure that can withstand them without abandoning the underlying thesis. Managing leverage and diversifying investments are therefore not simply ways to reduce losses. They buy investors the time needed to wait until the market ultimately validates their judgment, rather than forcing an exit at the worst moment.

Balancing insight into the future with lessons from past failures in risk management is what can turn a good investment idea into lasting performance.