Leverage thrills but certainty kills

Published in The Sunday Times on 20th September 2026.

You may have seen a recent story about a hedge fund blow up at the end of July. 

Leopold Aschenbrenner became prominent in technology circles after publishing a series of essays in 2024 arguing that rapid advances in artificial intelligence would require an enormous expansion of computing power, advanced semiconductors, memory and energy infrastructure. Those ideas, which have largely played out, became the intellectual foundation for his hedge fund, Situational Awareness. With no trading experience, the OpenAI alumnus became a Wall Street sensation over the last two years racking up reported gains in excess of 1,000%. 

Sharp declines in AI related stocks triggered margin calls in July, and his portfolio that had reportedly grown to tens of billions of dollars was forced into a fire sale. The fund was bought out of most of its public positions by Citadel, one of the largest hedge funds in the world. 

The real fuel was certainty

Almost all of the reporting focused on significant leverage used by the fund. But leverage was only the accelerant in this story. The real fuel was certainty.

Certainty leads investors to concentrate portfolios and abandon diversification – this is not necessarily a bad thing. It also encourages increased leverage – this can be a bad thing. It leads one to dismiss opposing views and ignore tail risks - definitely a bad thing. Many investment disasters begin not with a bad idea, but with a good idea taken too far.

The single most surprising thing to me in the last 30 years of working in financial markets, is how most people fear the wrong thing. Most investors think uncertainty is the enemy.

Uncertainty makes people nervous. It creates volatility. It fosters regret. As a result, many investors spend their lives seeking certainty. A sage. A forecast. A narrative that removes doubt. Most default to a low impact form of certainty by investing too much in assets – cash and bonds - that are subject to limited price variability. Notice I demurred from referring to these assets as low risk - they are only low risk if you are measuring risk incorrectly. 

Certainty changes behaviour

The danger isn't being wrong. In financial markets, that’s all but guaranteed. The danger is what happens when you become convinced you are right.

I’d a conversation with a friend recently about his pension. He works in the technology space and is convinced that AI is a bubble and that a massive correction is coming. With less than a 10% allocation to equities, he’s positioned for financial Armageddon.

It’s difficult to know where to start with this. Firstly, I have to acknowledge that he may be right. Severe bear markets, whilst rare, do occur. Typically, significant market drawdowns are the ones associated with recessions. There’s little current evidence to support that, but his argument is easier to dismiss than it is to refute. 

Never before has so much capital been committed to a single technological theme by such a small group of companies. If AI revenues fail to live up to the optimism embedded in current valuations, and if models become commoditised rather than differentiated, today's virtuous circle of spending, profits and rising share prices could quickly reverse.

If you’re looking for reasons not to invest – I’ve a long list I can provide. And there is a whiff of the dotcom era about this market. There are heroic assumptions embedded in the way that AI is being financed so it’s not hard to imagine a future in which expectations are severely disappointed. 

At its low point in 2009, the stock market had erased an entire decade of gains, trading below its 1999 level. Experiences like that certainly shaped me, leaving me predisposed to see threats and vulnerabilities everywhere.

There are myriad concerns about stock markets currently which are legitimate and deserve attention. But when I reflect back on my experience in financial markets, the cheerful optimist has had the bragging rights for the vast majority of the time. Being bearish all the time is easy – being bearish at the right time is something very different.

Excessive confidence in one world view

Both Aschenbrenner and my friend are making opposite mistakes born of the same source - excessive confidence in a single future. There is a version of the future in which the AI bubble bursts causing a severe bear market. But there are more future possible paths in which it doesn’t. 

There are ways to manage the risks my friend sees, without making a large bet on it. Geographically there are regions – Europe and the UK for example – which have much lower exposure to the AI theme. To be clear – such is the scale of the bet on AI, particularly in the US, that a collapse would have broad market implications. 

Investing in the stock market has been very rewarding over the last 10, 20 and 30 years. Those rewards come from bearing the uncertainty than others find intolerable. Stock market investors are paid to embrace uncertainty. The goal is to survive it, not avoid it. And you’ve a much better chance of that without leverage.

Gary Connolly is Investment Director at Davy. He can be contacted at gary.connolly@davy.ie or on X at @gconno1.