Wall Street analysts are now forecasting that global hyperscale AI capital expenditure (capex) could exceed $1 trillion in 2027, driven by massive spending increases from tech giants Microsoft, Amazon, Alphabet, and Meta. Combined spending for 2026 is estimated between $800 billion and $900 billion, representing a 67% year-over-year surge.
Following Q1 earnings calls, Microsoft raised its 2026 capex outlook to $190 billion, up 24% from a prior forecast of $154 billion. Amazon held firm at $200 billion, while Alphabet lifted its forecast by 4% to $185 billion. Meta increased its range to between $125 billion and $145 billion, up from $115–$135 billion. Notably, Microsoft attributed $25 billion of its increase to higher component costs, and Meta CEO Mark Zuckerberg cited memory pricing as a key driver. However, Meta's free cash flow dropped sharply to $1.2 billion in Q1 from $26 billion a year earlier, prompting Jefferies analysts to say Meta 'likely remains in the penalty box pending clearer capex ROI.'
On the positive side, Alphabet's cloud revenue surged 63% year-over-year, boosting its shares by roughly 10%, while its backlog grew 400% annually to $462 billion. Microsoft reported an annualized AI sales run-rate of over $37 billion, up 123% year-over-year, and Amazon's AWS posted its fastest growth in over three years at 28%, driven by AI workloads. Alphabet now processes more than 16 billion Gemini tokens per minute.
The spending boom is a clear tailwind for semiconductor companies. RBC Capital Markets highlighted Nvidia, Micron Technology, Marvell, Arm Holdings, and Astera Labs as well-positioned beneficiaries. Intel also posted strong Q1 earnings, and Evercore analysts pointed to growing demand for custom chips like TPUs, Trainium, and Maia processors, calling it a potential 'CPU renaissance.' AI demand is driving double-digit growth in wafer fab production, and supply for high-end AI compute is expected to remain tight throughout 2026.
