US Hyperscalers Set $720 Billion 2026 AI Capex as $300 Billion Boom Lifts Inflation
Updated
Updated · FinTech Global · Aug 7
US Hyperscalers Set $720 Billion 2026 AI Capex as $300 Billion Boom Lifts Inflation
3 articles · Updated · FinTech Global · Aug 7
Summary
$720 billion in 2026 capital spending is projected for the five largest US hyperscalers, extending an AI infrastructure surge that LSEG says has become a macroeconomic force.
$300 billion was already committed in 2025 by Amazon, Anthropic, Google, Meta, Microsoft, OpenAI and Oracle across chips, data centres, power and specialised labour, reinforced by fast-rising AI revenue such as Microsoft’s $37 billion run rate.
That spending is feeding through to prices: computer software and accessories rose nearly 14% over 12 months, while wholesale electronic components jumped 28%, reversing years of decline in some tech categories.
Trade and utilities are also absorbing the shock, with US computer imports more than doubling to about $93 billion in Q1 2026, semiconductor imports up 40%, and consumer electricity prices rising 4.6% as data-centre demand lifts power output.
LSEG says the current phase is inflationary but could turn disinflationary later if AI-driven productivity gains outweigh labour and capital costs, a transition that leaves central banks balancing near-term price pressure against longer-term efficiency gains.
Are flawed inflation metrics hiding the true economic impact of the multi-trillion-dollar AI infrastructure boom?
Will the massive energy demands of AI data centers ultimately price everyday consumers out of the electricity grid?
Could technical bottlenecks in advanced memory chips derail the tech industry's trillion-dollar AI expansion before productivity peaks?
The 2026 AI Capex Explosion: Hyperscaler Spending, Supply Chain Strains, and the ROI Dilemma
Overview
In 2026, surging demand for advanced AI models has driven the five largest U.S. hyperscalers to commit over $690 billion in capital spending, but this unprecedented investment faces immediate supply chain and energy bottlenecks. Severe power grid constraints and long equipment lead times mean up to half of planned data center capacity will be delayed, while geopolitical shocks have sent energy costs soaring for critical chipmakers. Despite this massive buildout, a fundamental lag in enterprise AI adoption has created a $500 billion revenue gap, fueling Wall Street skepticism and exposing the risks of circular financing and heavy leverage among new GPU cloud providers. Meanwhile, the rapid expansion of data centers is straining global resources, with electricity and water demands set to double by 2030, prompting urgent shifts toward more efficient AI architectures and hardware.