Big Tech is on track to spend $700 billion on AI infrastructure this year, with no clear end to the buildout in sight. The capex wave is reshaping power, real estate, semiconductor, and labour markets at a scale that has never been compressed into a single year.
The Numbers
- $700 billion in 2026 alone — combined hyperscaler AI capex
- Driven by Microsoft, Google, Amazon, Meta, plus Oracle, Apple, NVIDIA partners
- Year-on-year increase: roughly 40%+ over 2025
- Multi-year trajectory implies $2 trillion+ cumulative through 2028
Where The Money Goes
- Data centre construction — site, shell, fit-out
- GPU/TPU/ASIC procurement — NVIDIA, AMD, Google TPU, Amazon Trainium
- Power generation + transmission — direct PPA’s, on-site generation, nuclear partnerships
- Cooling + networking — liquid cooling, high-bandwidth interconnect
- Land acquisition — multi-decade commitments in select markets
The Macro Pressure
- Power-grid stress increasingly visible — utilities raising capacity tariffs
- Land prices in data-centre corridors (NoVA, Phoenix, Iowa, Texas) accelerating
- Skilled-labour wage inflation in electrical + HVAC trades
- NVIDIA + AMD revenue trajectories underwriting most of the capex curve
The Bear Case
- Replacement-cycle risk: GPU generations obsolete in 2–3 years
- Demand visibility beyond current AI use cases is contested
- Capex ROI math depends on continued enterprise AI adoption ramp
- Buffett’s “gambling” comment from Saturday’s AGM is the macro foil
Why It Matters
$700 billion in a single year is a structural reshaping of corporate capital allocation. It anchors NVIDIA’s earnings trajectory, drives utility-sector capex plans, props up real estate in select corridors, and absorbs trades-labour supply for the rest of the decade. The question is no longer whether the buildout happens — it’s whether the demand keeps pace.
Follow Vibes Uncut Media for continuing AI capex coverage.
Leave a Reply