Construction Criteria

The AI Infrastructure Portfolio was built around the goal of capturing the most direct and sustainable beneficiaries of long-term artificial intelligence adoption. Rather than focusing on software or short-term speculative names, the portfolio emphasizes the underlying hardware and infrastructure that enable AI computation, storage, and deployment at scale.

1. Core Theme Alignment
Each company included must play a critical role in the AI value chain, from semiconductor fabrication and chip design to data center networking, memory, power, and cooling. The focus is on measurable exposure to AI-related revenue growth and capital investment, not secondary or peripheral beneficiaries.

2. Proven Execution and Financial Health
Holdings are screened for strong balance sheets, consistent free cash flow, and established competitive positions within their sectors. Preference is given to companies with high return on equity, moderate debt levels, and resilient margins through market cycles.

3. Sector Diversification
To reduce volatility and improve long-term stability, the portfolio blends multiple sectors tied to AI infrastructure:

• Semiconductors and Fabrication: NVIDIA, TSMC, ASML, Micron
• Networking and Connectivity: Broadcom, Arista Networks
• Power and Cooling Infrastructure: Vertiv, Eaton

4. Market Leadership
Each position represents a recognized leader in its respective segment, measured by market share, innovation leadership, or long-term pricing power. The intent is to maintain exposure to firms with durable competitive advantages rather than emerging or unproven players.

5. Data-Driven Weighting
Portfolio weights are determined based on a mix of company size, historical performance, and risk-adjusted return potential. Larger, more established companies with proven growth records, such as NVIDIA and TSMC, receive higher allocations, while smaller or more cyclical firms hold lower weights.

6. Long-Term Orientation
The portfolio is designed for multi-year observation rather than active trading. Selections are meant to represent the structural foundations of AI infrastructure over the next decade, aligning with the belief that hardware, manufacturing, and energy systems will remain the key bottlenecks, and therefore the primary value drivers of AI expansion.