The AI industry has been on a wild ride since its inception, with many investors and market participants wondering when the party will end. While some segments of the market, such as semiconductors, have seen extraordinary gains, other areas like software and internet remain strong funding sources. The question on everyone’s mind is no longer whether AI is real but when the trade will start to cool off.

Investors are eagerly searching for early warning signs that the AI boom may be reaching its peak. Two key indicators are closely watched: the growth trajectory of major AI labs and scaling laws. The industry’s spending ambitions have been fueled by the assumption that revenues will continue to compound at remarkable rates, but any significant slowdown in this growth could raise questions about the sustainability of the current spending cycle.

Major model upgrades are expected before year-end, and expectations are higher than ever. The upgrades not only need to be good but must justify one of the largest capital spending cycles in market history. JPM’s Mark Schilskey notes that the market is effectively betting on larger models, more compute, and next-generation hardware delivering meaningful performance improvements. However, we have yet to see fully Blackwell-trained models deployed at scale, and if the next wave of models does not deliver another step-change in capabilities, investors may start questioning whether the industry’s spending ambitions are ahead of reality.

The AI boom has been a remarkable phenomenon, but it is important to recognize that no market remains bifurcated forever. As always, it is crucial to stay vigilant and monitor key indicators to ensure a continued successful investment strategy in the AI space.