The Tech Download: Can hyperscalers justify their huge AI capex? (www.cnbc.com)

Hyperscalers like Amazon, Microsoft, Meta, and Alphabet are projected to announce a staggering $700 billion in capital expenditures (capex) this year, driven by the unprecedented boom in AI. This massive investment reflects the urgency for these tech giants to expand their data center capacities to support AI technologies. While many analysts remain bullish on hyperscaler stocks due to the pre-sale of data center capacity, investors are uneasy, particularly after a recent market plunge that wiped over $1 trillion from Big Tech valuations. Analysts express that the current level of capex is unsustainable, consuming nearly 100% of operational cash flow, compared to a historical average of 40%. The implications of such spending raise critical questions for the AI/ML community. The potential returns from these investments rely heavily on timely monetization of AI services, with a pressing need for hyperscalers to demonstrate clear pathways to profitability. With much of this infrastructure having an estimated useful life of just 3-5 years, the timeline for recouping these substantial expenditures is tight, creating a "binary" situation where success hinges on rapid demand for AI solutions. Until hyperscalers provide credible strategies to assure investors of significant future returns, market volatility is likely to persist.