US tech giants and enterprises face rising AI infrastructure spending with unclear returns, pressuring profitability and investor sentiment.
The four largest US tech firms—Alphabet, Amazon, Meta, and Microsoft—are projected to spend $650 billion on AI infrastructure by 2026. Analysts forecast total hyperscaler AI capex could exceed $1 trillion in 2027, signaling a rapid escalation in costs.
Enterprise AI spending is also climbing sharply. Average monthly AI costs for software companies rose 36% year over year to $85,521, while firms spending over $100,000 monthly nearly doubled to 45%. Yet only 51% of organizations can measure returns on this spending.
The shift from growth focus to cost scrutiny highlights structural challenges in AI economics, where inference costs per query add ongoing expenses absent in traditional software models.