The Tech Download: Can hyperscalers justify their huge AI capex?

In a rapidly evolving landscape driven by artificial intelligence, major tech companies are making staggering investments to expand their data center capabilities. Hyperscalers such as Amazon, Microsoft, Meta, and Alphabet have committed to capital expenditures that could soar to $700 billion this year—an amount that surpasses the GDP of nations like the United Arab Emirates, Singapore, and Israel. This aggressive spending spree has left investors on edge. Recently, the stock prices of these Big Tech giants plummeted, erasing over $1 trillion from their market valuations due to rising concerns about the sustainability of such immense financial commitments. Although there has been a slight recovery in the market, apprehension regarding the future of AI spending continues to loom large. According to Michael Field, chief equity strategist at Morningstar, the anticipated capital expenditure for 2026 reflects a staggering 60% increase compared to the previous year. He warns that this situation could become precarious: "At a certain point, this bet becomes binary: either demand and monetization will follow and justify the expenditure, or it won’t, leading to business failures," he stated. Investors previously felt secure with these investments being a mere side gamble, but the current stakes have raised their level of discomfort significantly. UBS analysts report that this year's capital expenditure from hyperscalers is set to consume nearly 100% of their operational cash flow, in stark contrast to a decade-long average of just 40%. Bob Savage, head of markets macro strategy at BNY, emphasized that the concern lies not merely in the expenditure itself, but in the financing methods. He noted, "If this increases the net borrowing of mega-caps, it detracts from equity holdings." Notably, Oracle has disclosed plans to secure between $45 billion and $50 billion in capital over the next fiscal year, while Alphabet aims to raise $20 billion through a U.S. dollar bond sale, as reported by CNBC. Despite the market's volatility, many analysts maintain an optimistic outlook regarding hyperscaler stocks. Gil Luria, head of technology research at D.A. Davidson, pointed out that leading data center builders like Amazon, Microsoft, and Google are already seeing positive returns as they pre-sell their capacity. He anticipates that as AI usage expands, both consumers and businesses will be willing to invest more in the value generated, leading to further profitability. However, the timeline for recouping these massive investments remains uncertain. Field cautioned that the useful lifespan of much AI-related spending, including data centers and chips, could be as short as 3-5 years. This tight window necessitates significant returns on investment before 2030, making credible strategies for monetization and clear payback timelines essential for reassuring investors. As Alphabet returns to the debt market to support its AI initiatives, the landscape continues to shift. Amidst the backdrop of these developments, geopolitical tensions are also escalating, raising questions about digital sovereignty, particularly in relation to the semiconductor supply chain. The unfolding scenarios present a complex picture for both investors and the tech industry as a whole.