Capex Surge Raises Questions About AI Economics
Microsoft set the tone last week when it disclosed $34.9 billion in quarterly capital expenditure, with almost half funnelled into GPUs and CPUs to power AI systems. Alphabet plans to push its 2026 capex beyond $90 billion, Meta expects as much as $72 billion, and when Amazon is added to the mix, the combined number crosses $380 billion.This unprecedented spending spree has set off a fresh wave of concern.
Investors fear that while the long-term AI opportunity remains intact, the near-term economics look strained. If revenue growth does not keep pace with rising infrastructure costs, margins could come under pressure. In a market where these stocks already trade at elevated valuations, even a small disappointment can spark large corrections.Market Concentration Risk Comes Back Into Focus
The growing dominance of the Magnificent 7 has intensified market vulnerability. With more than a third of the S&P 500 now tied to just seven companies, any shift in AI sentiment disproportionately impacts the broader index.
AI Enthusiasm Moderates, Not Disappears
Despite the churn, analysts say investors are not walking away from AI. Instead, the tone has shifted from unbridled optimism to selective allocation. The long-term commercial potential of AI remains strong, but the path to monetisation is uncertain and capital-intensive.
The emerging “circular economy” around AI — where Big Tech companies buy GPUs, lease cloud capacity and sign multi-year compute agreements largely within their own ecosystem — is also raising eyebrows. Multi-year supply contracts with NVIDIA further lock in spending, raising the question of whether growth is demand-driven or market-structure driven.
