Is the AI Investment Boom Built on a Solid Foundation?

Let’s talk tech. You know how everyone is obsessed with AI right now, especially the "Magnificent Seven" companies? This enthusiasm has created an enormous capital expenditure (CapEx) cycle, often called the AI CapEx cycle . These massive corporations are generating incredible operating profit margins, averaging around 30% for the biggest players . When you can borrow money at, say, 4% and reinvest it to earn a 30% return, you would be foolish not to borrow and expand your facilities aggressively.

This enormous injection of capital into new equipment and technology is quickly capitalized on the balance sheet, meaning it turns into a long-term asset . Here's where the risk comes in: assets must be depreciated over time, generating a non-cash expense . If the pace of CapEx is too hot, the resulting depreciation expense in the coming years will be massive . I've found that investors often overlook this accounting consequence during the excitement phase of a boom. For these heavy depreciation costs to be absorbed without hurting profitability, the companies need to generate equally huge returns and productivity gains from that equipment.

The surprising insight here is that the sheer size of the debt and investment (CapEx) makes the entire sector highly sensitive to any doubt about growth . If the growth narrative falters even slightly, these stocks—which are priced for near-perfection—could violently correct, not necessarily because the core technology is bad, but because the financing structure has become precarious. While the long-term promise of AI is undeniable—it’s a technology that must be included in any well-diversified portfolio—the current intensity of investment and the resulting debt load constitute a significant, often underappreciated, risk that warrants caution when considering concentrated positions.