The Capex Multiplier — From Cloud Spend to Equipment Orders
Understanding the flow from cloud providers to tool makers.
The Chain
Hyperscalers
(Google, Microsoft, Amazon, Meta)
↓
Cloud Infrastructure
(Data centers, servers, networking)
↓
AI Chips
(NVIDIA, AMD, Broadcom, Marvell)
↓
Foundries
(TSMC, Samsung, Intel)
↓
Wafer Fab Equipment (WFE)
(ASML, LAM, AMAT, KLA)
The Numbers
| Level | Example Spend | Key Suppliers |
|---|---|---|
| Hyperscaler Capex | $180B (Google 2026e) | Dell, HPE, NVIDIA |
| AI Chip Revenue | ~$200B industry | NVIDIA, AMD, Broadcom |
| Foundry Capex | $50-60B (TSMC 2026e) | Equipment makers |
| WFE Revenue | $130B (2026e) | ASML, LAM, AMAT, KLA |
The Rule of Thumb
- Hyperscaler capex grows → AI chip demand grows → Foundry utilization rises → WFE orders accelerate
- Lag time: 6-12 months from hyperscaler guidance to WFE bookings
- Amplification: A 10% move in hyperscaler capex can drive 15-20% move in WFE
Case Study: 2023-2024 AI Boom
- ChatGPT launches (Nov 2022)
- Hyperscalers announce massive AI capex (2023)
- NVIDIA orders surge at TSMC (mid-2023)
- TSMC capex rises; CoWoS bottleneck emerges
- WFE orders accelerate for advanced packaging (late 2023-2024)
Key Insight
Watch hyperscaler capex guidance in earnings calls. It's the leading indicator for the entire chain.