Beyond Cap Rates and NOI
Traditional real estate underwriting metrics — cap rates, NOI multiples, price per square foot — are necessary but insufficient for data center acquisitions. Infrastructure assets require a deeper technical and market evaluation.
The Five-Layer Evaluation Framework
We recommend evaluating data center acquisitions across five dimensions:
1. Power Infrastructure
- Utility capacity — what’s the total available utility feed, and how much headroom exists?
- Redundancy architecture — N, N+1, 2N? Does it match target customer requirements?
- Power cost structure — fixed vs. variable rates, demand charges, renewable energy availability
- Upgrade path — can utility capacity be expanded, and at what cost and timeline?
2. Connectivity Position
- Carrier density — how many network providers serve the facility?
- Cloud on-ramps — direct connections to AWS, Azure, GCP within the facility or metro
- Fiber diversity — multiple fiber paths and providers for redundancy
- IX presence — internet exchange access for peering and traffic exchange
3. Market Dynamics
- Supply pipeline — what new capacity is under construction or planned in the market?
- Demand drivers — enterprise, cloud, AI/ML, government — what’s driving absorption?
- Pricing trends — are rates stable, rising, or under pressure?
- Competitive set — who else operates in the market and at what scale?
4. Physical Plant
- Age and condition — mechanical and electrical systems lifecycle assessment
- Density capability — can the facility support high-density AI workloads (30kW+ per rack)?
- Expansion potential — is there room to grow on the existing site?
- Environmental compliance — any legacy issues, regulatory constraints, or remediation needs?
5. Customer Profile
- Tenant concentration — revenue concentration risk across the customer base
- Contract terms — remaining lease duration, escalation structures, renewal options
- Credit quality — tenant financial strength and payment history
- Churn risk — likelihood of tenant departure at lease expiration
Common Mistakes
The most frequent errors we see in data center acquisition evaluation:
- Overvaluing square footage — space without power is worthless in today’s market
- Ignoring connectivity — a facility with limited fiber options has a ceiling on its value
- Underestimating capex — retrofit costs for AI-ready infrastructure are consistently underestimated
- Missing market timing — buying into a market just as new supply is delivering
The Stacked AI Approach
We help capital partners navigate these evaluations with market-specific intelligence, technical due diligence support, and deal structuring expertise. Every acquisition is different, but the framework for evaluation should be consistent and rigorous.
The best data center acquisitions aren’t found — they’re sourced through relationships and evaluated through deep market knowledge.