🖥️ Downstream Products

Semiconductors are useless without products that consume them. The downstream trade captures the infrastructure and applications that turn chips into value: data centers housing AI servers, cloud platforms running on those servers, and robots powered by advanced processors. When semiconductor demand surges, it’s because these downstream segments are spending.

Why downstream matters for semiconductor traders

The relationship between chips and their consumers is bidirectional:

• Semiconductors enable data centers, cloud, and robotics
• Downstream capex drives semiconductor demand

Understanding downstream spending helps you:

1. Validate semiconductor demand (is it real or hype?)
2. Identify where chip demand is concentrated
3. Find correlated trades that benefit from the same cycle

The downstream value chain

  flowchart TD
    A[**Semiconductors**<br/>SMH, SOXX] --> B[**Data Center Infrastructure**<br/>DTCR]
    B --> C[**Cloud Computing**<br/>CLOU, WCLD]
    C --> D[**AI/Robotics Applications**<br/>BOTZ, ROBO]
    
    A -.- E[The chips]
    B -.- F[The buildings & power]
    C -.- G[The software platforms]
    D -.- H[The end use cases]

Each layer depends on the one above — and capital flows down the chain.

Data Center & Infrastructure ETFs

Data centers are the physical manifestation of AI demand. Every NVIDIA GPU needs a rack, power, cooling, and connectivity. This infrastructure buildout is a multi-decade megatrend.

DTCR — The Data Center Play

What DTCR holds:

• Data center REITs (Equinix, Digital Realty, QTS)
• Tower companies (American Tower, Crown Castle)
• Fiber/network infrastructure

Key characteristics:

• REIT exposure: Many holdings are REITs with income characteristics
• Lower beta than semiconductor ETFs
• Beneficiary of AI capex without direct chip exposure

Why data centers matter

The AI revolution has created unprecedented data center demand:

Power is the constraint: AI data centers consume 10-30x more power than traditional facilities. This is creating bottlenecks in power availability and utility partnerships.

Data center REITs as signals

Data center REITs report:

• Leasing activity: Forward demand indicator
• Pricing/rent growth: Supply/demand balance
• Capacity additions: Capex flowing to infrastructure

When Equinix or Digital Realty reports strong leasing, it confirms semiconductor demand is translating to physical buildout.

Cloud Computing ETFs

Cloud computing sits between physical infrastructure and end applications. The hyperscalers (AWS, Azure, GCP) are the largest semiconductor customers — their capex directly drives chip demand.

CLOU — The Cloud Computing ETF

What CLOU holds:

• Cloud infrastructure (AWS exposure through AMZN, Azure through MSFT)
• SaaS companies (Salesforce, ServiceNow, Workday)
• Cloud security (CrowdStrike, Zscaler)

Key characteristics:

• Software-heavy: More software exposure than hardware
• Higher valuations: Growth stocks with premium multiples
• AI beneficiary: Cloud platforms deliver AI services

WCLD — The Pure SaaS Play

What WCLD holds:

• Pure-play SaaS companies
• Emerging cloud businesses (smaller-cap focus)
• Excludes mega-cap hyperscalers

CLOU vs WCLD:

Hyperscaler Capex as a Signal

The three major hyperscalers announce capex quarterly:

Rising hyperscaler capex = bullish for semiconductors. These are the customers buying NVIDIA GPUs by the hundreds of thousands.

Robotics & AI ETFs

Robotics and AI represent the end applications of semiconductor technology. These ETFs capture companies building products that consume chips at scale.

BOTZ — The Robotics ETF

What BOTZ holds:

• Industrial robotics (Fanuc, ABB, Keyence)
• Automation equipment (Rockwell, Cognex)
• AI companies (NVIDIA appears here too)
• Autonomous vehicles (some exposure)

Key characteristics:

• Industrial tilt: Heavy manufacturing automation
• Japan exposure: Many robotics leaders are Japanese
• Cyclical: Tied to manufacturing capex

ROBO — The Broader Automation Play

What ROBO holds:

• Broader automation exposure than BOTZ
• Healthcare robotics (surgical robots)
• Logistics automation (warehouse robots)
• Consumer robotics

BOTZ vs ROBO:

ARKQ — The Disruptive Automation Play

What ARKQ holds:

• Tesla (typically largest holding)
• Autonomous vehicles
• 3D printing
• Energy storage

Key characteristics:

• Actively managed by ARK Invest
• High conviction: Concentrated positions
• Volatile: ARK’s style creates big swings

How do downstream ETFs relate to semiconductors?

The Demand Confirmation Chain

Healthy semiconductor rally: All downstream segments participating Suspicious rally: Chips up, downstream flat (demand not confirmed)

Key Ratios to Monitor

Signal Interpretation

Cycle dynamics

Early Chip Cycle

• SMH leads everything
• Downstream ETFs lag
• Skepticism about AI capex
• Data center REITs consolidating

Mid-Cycle

• Hyperscaler capex announcements confirm demand
• DTCR starts catching up
• CLOU participation broadens
• Robotics/automation orders rising

Late Cycle

• Downstream fully participating
• Data center supply concerns (too much capacity?)
• Cloud growth rates moderating
• Watch for capex guidance cuts

Risk considerations

Different Risk Profiles

Correlation with Semiconductors

Downstream ETFs are correlated but not identical to semiconductors:

• DTCR: Lower correlation, REIT characteristics
• CLOU: Moderate correlation, software dynamics
• BOTZ/ROBO: Moderate correlation, industrial cycle overlay

This makes them useful for diversification within the tech theme.

When to use downstream ETFs

Scenarios Favoring Downstream

Scenarios to Be Cautious

Quick reference

Related themes

Downstream semiconductor consumption connects to energy and materials:

Sources