Bitcoin has crashed. Could AI be next?
The father of a field of mathematics known as fractal geometry, Benoit Mandelbrot, in his fantastic book, The (Mis)behaviour of Markets, observed that market price movements, unlike the smooth curves assumed in traditional financial models, are jagged and irregular, meaning that large price swings are more frequent than expected, and periods of high volatility can be followed by more high volatility. His work provides a more realistic framework for understanding and modeling risk, noting also that volatility in markets tends to cluster.
In my experience, I’ve observed those volatility clusters often mark turning points in markets. To be clear, not all periods of heightened volatility are turning points, but almost all turning points are accompanied by greater volatility.
When this article was written (on 21 November 2025) we experienced some of that rising volatility. Could it be followed, as Mandelbrot observed, by even more in the near future?
Nvidia, the undisputed leader in the artificial intelligence (AI) boom, delivered another stellar earnings report, and after the bell, on 19 November 2025 – beating expectations with massive revenue growth, sky-high profits, and guidance that once again pointed to explosive demand ahead.
Initially the market reacted positively. The stock gapped up sharply at the open, climbing more than 5 per cent in the early hours as investors piled in, betting this would ignite a fresh leg higher for tech and the broader market.
But then… something shifted.
By the close, all those gains had vanished. Nvidia flipped red, finishing down more than three per cent, dragging the Nasdaq and S&P 500 lower with it.
Risk assets took a hit across the board – Bitcoin, often seen as a proxy for investor appetite for high-growth plays, plunged nearly five per cent in a single day. Bitcoin is now 30 per cent lower than its high just six weeks ago. That’s a crash by anyone’s measure.
Peak AI? For now.
The moves don’t appear to be just random volatility. Moments like these – when a company reports phenomenal numbers that exceed expectations and raises its outlook, yet the stock (and the sector) sell off anyway – are warning signs market veterans pay attention to.
I’ve long spoken about the caution investors need to apply to AI-related stocks. As the boom took hold, I reminded our investors that course-of-history-changing tech has historically been better for consumers than investors. That was true of the internet boom of 1999/2000 – we all benefited enormously from the advent of the internet, but internet stock investors in 1999 and early 2000 lost billions.
Many pointed to the combination of Nvidia’s market-beating results and poor share price reaction as a potential turning point.
When ‘perfect’ results from the sector’s star name no longer sparks sustained buying and instead prompts a “sell the fact” response, it’s often a sign sentiment is near or at a peak. For now.
As we have noted over the last few weeks, the risk is rising that the market will cotton on to the idea that, as the AI boom approaches the reality of selling more AI software to a ‘cyclical’ consumer, the AI theme narrative itself cannot continue to be accepted as ‘structural’.
Consumers – those individuals and businesses that must buy artificial intelligence (AI) products to generate acceptable returns for the hyperscalers to justify the trillions they’re spending on the AI infrastructure – are cyclical. They might not be able to buy enough AI software to make the numbers stack up. We have recently offered a few calculations to say, they won’t. They can’t.
And maybe, just maybe, the rest of the market is catching on. That’s why sentiment appears to have shifted.
Is the AI Boom over?
Does (last) week’s price action mean the AI boom is over? Are we at the absolute top of this multi-year rally in tech? Probably not. But who knows? We can’t declare that after one day’s price action. Remember, markets can stay irrational longer than most expect, and Nvidia’s fundamentals are incredibly strong. It’s likely this isn’t the peak of the AI boom.
Thursday’s (20 November 2025) sharp intraday reversal, however, shouldn’t be ignored. It suggests that beneath the surface, investors are questioning how much more upside is already priced in, especially with valuations stretched and broader economic uncertainties lingering.
It’s the first shot over the bow of the boat. The first crack in the polished concrete that is the foundation of the AI boom.
What to do now?
If you’re heavily allocated to AI themes, and AI-related growth stocks, this is a good time to step back and stress-test your holdings. Ask yourself: How would my portfolio hold up if we see a meaningful pullback or rotation out of these high-fliers in the coming weeks or months?
Some investors have rebalanced, reducing exposure to their AI winners and index exchange traded funds (ETFs) like the S&P 500, which has a huge weighting to the AI hyperscalers, and applied for AA-rated private credit and market-neutral funds. Some have raised their cash weighting a little. Berkshire Hathaway currently has a record US$381.7 billion in cash or approximately 53.6 per cent of its portfolio.
Few financial advisers would suggest investors hold that much cash, often because it would inspire tax consequences, and add the burden of re-entering the market at the right time.
Nevertheless, a prudent review now could go a long way toward protecting gains you’ve enjoyed during this extraordinary run.
(Please note that this article was written on 21 November 2025, so all prices and movements are related to this date.)