Episode Details

Back to Episodes
Hyperscalers Push AI Capex Toward $630B as Grid Rules Loom

Hyperscalers Push AI Capex Toward $630B as Grid Rules Loom

Episode 11 Published 1 month, 3 weeks ago
Description

Amazon, Google, Meta and Microsoft are pointing to as much as $630 billion in 2026 capex, while OpenAI’s Stargate and FERC’s large-load rulemaking show the AI infrastructure race moving from GPU orders into grid jurisdiction and site capacity.

In this episode

  1. Hyperscalers Plan $630 Billion in 2026 CapEx — Datacenter Richness

    # Hyperscalers Plan $630 Billion in 2026 CapEx ### Investment in GPUs, Data Centers for AI Poised to Surge 62% From 2025 Record [Rich Miller](https://substack.com/@richmiller) Feb 06, 2026 ∙ Paid An Amazon Web Services data center. Amazon said Thursday that it expects capital expenditures of $200 billion this year for AI infrastructure. (Photo: Amazon) Despite historic investments in GPUs…

  2. Step Back — FERC is being pushed to write rules for large-load interconnections, but load has usually been a utility and state-commission problem—what’s the actual jurisdictional hook here, and where could a federal rule meaningfully speed up or reshuffle AI data center projects versus just adding another process layer?

    Background sources

  3. Hyperscale Spending Spree is Driving Dramatic Growth in Data Ce
Listen Now