Major technology companies, including Amazon, Google, Meta, and Microsoft, are significantly escalating their capital expenditures on AI infrastructure, primarily for building data centers and acquiring advanced chips. Over the second quarter, capital expenditures for these four companies collectively reached $170 billion, marking a 72% increase year-over-year. Projections indicate that their combined spending for this year and next could hit a staggering $1.5 trillion, a figure that continues to rise with each quarterly estimate. This aggressive spending is driven by the race to establish more data centers and the soaring prices of essential components like memory chips, with Microsoft alone anticipating an additional $25 billion due to component price hikes.
This unprecedented investment spree is impacting the financial health of some tech giants. Google, for the first time since its 2004 IPO, reported negative free cash flow, indicating that its operational and infrastructure building costs exceeded its revenue, causing its stock to fall over 6%. Similarly, Meta saw its stock drop more than 7% after disclosing that its AI outlays were growing faster than its revenue, and its cash balance declined 91% year-over-year by June. Microsoft's cash reserves also decreased by 23% to $19.6 billion. Legacy Capital analysts and others are raising alarms about when these substantial expenditures will yield justifiable returns.
Despite investor concerns, companies like Alphabet's finance chief, Anat Ashkenazi, insist they will continue to invest as long as attractive opportunities exist. The heavy investment is creating a massive backlog of signed contracts for Amazon, Google, and Microsoft, totaling nearly $1.7 trillion, more than double the previous year. However, Moody's Ratings points to a concentration risk, as much of this backlog stems from partnerships with leading AI startups like OpenAI and Anthropic, raising questions about whether these startups will truly need all the computing power requested and their ability to pay for it. Goldman Sachs estimates that AI capital expenditure now consumes 93-94% of hyperscaler operating cash flow, a sharp increase from 33-40% in 2022-2023. Analysts from Itaú BBA also note that memory chip inflation is impacting Apple's margins, as suppliers prioritize data center supply chains over other sectors.
