Wall Street is currently locked in a tense debate over whether artificial intelligence is a sustainable revolution or a classic speculative bubble. While most analysts focus on whether the massive capital expenditures of tech giants will eventually yield proportional profits, Dario Perkins of TS Lombard warns that the real danger might come from directions investors aren’t even looking. According to Perkins, the catalyst for a market correction does not necessarily have to originate from within the technology itself; instead, external shocks could easily pull the rug out from under the current euphoria.
A primary concern is that much of the perceived growth in AI revenue is actually just circular spending within its own ecosystem. Rather than seeing widespread adoption where general consumers and diverse businesses pay for productivity gains, there is evidence of capex recycling where AI companies simply sell to one another. This fragility is compounded by a growing mountain of debt among smaller cloud providers and infrastructure firms who lack the fortress balance sheets of titans like Microsoft or Google. With estimated debts reaching two trillion dollars by 2027, any sudden dip in demand or a wave of cancelled orders could create a destructive ripple effect throughout the supply chain.
Beyond the balance sheets, Perkins points to the volatile nature of hardware cycles and the looming threat of a semiconductor glut. Because companies often over-order chips to avoid shortages, the industry is susceptible to a bullwhip effect where an abrupt slowdown leads to catastrophic excess supply. Higher interest rates only make this environment more precarious for those relying on borrowed money to scale their operations during this gold rush.
Perhaps most unexpectedly, some of the greatest risks are sociological rather than financial. While enthusiasts celebrate AI progress on social media, there remains significant public resentment regarding job losses and the environmental impact of sprawling data centers. Perkins suggests that voters possess power that investors often ignore, noting that a single high profile disaster or accident could suddenly turn AI toxic in the eyes of the public. Such a shift in sentiment could lead to rapid political blowback and regulatory crackdowns that would dismantle the boom faster than any quarterly earnings report ever could.
