Particle.news
Download on the App Store

AI Spending Surges Into 2026 as Nvidia Signals Relentless Demand

Energy availability, depreciation costs, capital recovery now define the economics of the AI data‑center buildout.

Overview

  • Nvidia guided to about $65 billion in fiscal Q4 revenue and said it has visibility to roughly $0.5 trillion of Blackwell and Vera Rubin orders through 2026, reinforcing its AI‑accelerator lead as reports note U.S. approval to sell H200 chips in China.
  • Big Tech is preparing higher outlays after roughly $440 billion was flagged for the next year, with Microsoft disclosing $34.9 billion in capex in one quarter and Alphabet saying its 2025 spend rose to about $85 billion with further increases planned for 2026.
  • The International Energy Agency projects data‑center electricity use could climb from about 415 TWh in 2024 to nearly 945 TWh by 2030, raising questions about grid access, shorter hardware lifecycles, and whether long‑lived assets will earn adequate returns.
  • Venture funding remains concentrated in AI as reports show more than half of VC dollars and 36% of deals went to the sector in 2025, with investors expecting brisk 2026 activity across chips, data centers, robotics, and AI applications.
  • Debate over valuations and sustainability is intensifying as Thiel Macro exited Nvidia and cut Tesla while building Microsoft, even as Wedbush’s Dan Ives projects a 20%–25% tech advance in 2026 on expanding AI monetization.