A trillion dollars is about to pour into AI infrastructure in 2026 and most of the winners aren't the names you think.
Milk Road AI Twitter · Milk Road AI (@MilkRoadAI) · 2026-08-03
Goldman Sachs projects $1.019 trillion in global AI infrastructure capex for 2026, with physical infrastructure suppliers like GE Vernova, Quanta Services, and Comfort Systems positioned to benefit alongside chip makers as hyperscalers each commit over $100 billion.
Appears in
Extraction
Topics: ai-infrastructurecapital-expenditureenergy-infrastructuresemiconductorsinvestment-analysis
Claims
- Goldman Sachs estimates total global AI-related capex at $1.019 trillion in 2026, with $581 billion originating from the US.
- Google, Amazon, Microsoft, and Meta each plan to spend over $100 billion individually on AI infrastructure in 2026.
- GE Vernova booked $2.4 billion in data center electrification orders in a single quarter, exceeding all of 2025 combined.
- Quanta Services holds a record $48.5 billion backlog tied directly to grid construction for data center power delivery.
- Physical infrastructure suppliers for power, cooling, and networking benefit from AI buildout regardless of which hyperscaler dominates.
Key quotes
A trillion dollars is about to pour into AI infrastructure in 2026 and most of the winners aren't the names you think.
GE Vernova...booked $2.4 billion in electrification orders for data centers in a single quarter, more than all of 2025 combined, with shares up nearly 180% over the past year.
Every new gigawatt of AI capacity needs power, cooling and networking regardless of which hyperscaler is building it.