Investing.com — A new industry report from BofA Global Research suggests that aggressive investments in artificial intelligence (AI) could drive global hyperscale capital expenditure (capex) to $1 trillion by calendar year 2027.
Analysts note that major U.S. technology firms are significantly raising their spending outlooks as AI sales accelerate and supply constraints for high-end compute hardware persist through 2026.
The upward revision follows first-quarter earnings reports from leading hyperscalers, including Alphabet (GOOGL), Microsoft (MSFT), Amazon (AMZN), and Meta Platforms (META).
BofA now forecasts that combined hyperscale capex will exceed $800 billion in 2026, a 67% year-over-year increase, before crossing the $1 trillion threshold the following year.
The surge in spending is backed by rapidly growing AI revenue streams across the major cloud providers. Alphabet is now generating over 16 billion Gemini tokens per minute, with its search segment growing 19% due to AI-integrated queries.
Meanwhile, Microsoft reported an annualized AI sales run-rate exceeding $37 billion, which marks a 123% increase year-over-year. Amazon’s AWS has also seen its fastest growth in over three years at 28%, driven largely by AI workloads and strategic partnerships.
To meet the rising demand, the tech firms have substantially increased their 2026 capex guidance.
Microsoft raised its outlook to $190 billion, up from a prior market expectation of $154 billion, while Amazon maintained a strong guidance of $200 billion. Alphabet and Meta also adjusted their projected spends upward to $185 billion and $135 billion, respectively.
A critical component of the increased capex is the rising cost of hardware, which the report indicates is being absorbed by the hyperscalers. Microsoft specifically noted that $25 billion of its 2026 capex increase is attributed to higher component pricing.
The current environment grants significant pricing power to semiconductor vendors, who can pass on rising costs for wafers, memory, and substrates to their customers.
Key beneficiaries of the sustained spending cycle include compute giants like Nvidia (NVDA), as well as firms specializing in memory, optics, semicaps, and power semiconductors required to support massive data center expansions.
The report highlights that hyperscalers are currently placing equal emphasis on both merchant GPUs and custom silicon deployments.
Supply for AI compute is expected to remain tight throughout 2026, and the data suggests that strong customer commitments and improving free cash flow will continue to justify the historic levels of investment.
