Big Tech earnings season and the capex spiral

Big Tech earnings season begins next week with Google leading the charge. While investors usually focus on profit and revenue figures, the most critical data point this time centers on capital expenditure for AI infrastructure. Companies including Amazon, Microsoft, and Meta have committed over 700 billion dollars to this sector for the current year. The primary concern is whether these massive outlays translate into meaningful growth or simply cover the rising costs of necessary equipment.

Building AI data centers has become significantly more expensive. Prices for memory chips have increased, while access to power equipment, construction materials, and qualified labor remains restricted. Analysts note that the cost to construct one gigawatt of AI capacity has climbed by approximately 20 percent. A standard Nvidia-based configuration that previously cost 29 billion dollars now requires 35 billion dollars. This price surge creates a difficult cycle where increasing demand leads to further shortages and higher costs for everyone involved.

Industry research indicates that a substantial portion of the recent increase in capital spending serves to offset inflation rather than fund new capacity. Estimates suggest that 20 to 30 percent of future budget hikes will likely go toward higher costs. Investors are urged to look beyond total spending numbers to determine if these investments represent actual expansion. If companies report increased budget allocations without corresponding data on GPU deployments or new power capacity, those figures may indicate stagnant growth disguised by higher market prices.