Economic Availability Score
Pre-calculated from publicly available data. No access to your systems required. Select your operator or enter your own numbers.
How this works
Major data center operators publish Power Usage Effectiveness data in annual ESG and sustainability reports. Synestra uses that data to estimate the Economic Availability gap — the difference between what a campus generates and what it delivers to useful compute. The calculation is transparent. The sources are cited. The number is yours to test.
Select a perspective
Data center operators recover value through PUE reduction, BTM dispatch optimization, and stranded capacity. Tenants recover value through compute yield, GPU utilization, and workload efficiency. Select your perspective to see the relevant EA estimate.
Based on publicly available ESG data, most major AI infrastructure operators estimate a Full EA Score in the 50 – 56 range — Below Average to Underperforming.
Synestra's first-deployment target is 73+, moving operators to Competitive. Every point of improvement is a point of cost advantage in the AI infrastructure market.
The following EA estimates are derived from each operator's publicly disclosed Power Usage Effectiveness (PUE). Sources are cited below each result. These are scenario estimates, not measured customer outcomes.
EA Estimate
Data source
The cost of a score of 55
The numbers below are the energy floor only — the minimum verifiable cost of operating at the PUE levels major operators disclose publicly. The full EA opportunity is substantially larger once recovered compute capacity is included.
500 MW campus · PUE 1.28 · $40 / MWh · PUE gap 0.08
$1.15M per month · $14.0M per year · energy savings only · based on published PUE
Energy floor assumes PUE 1.28 → 1.20 (gap of 0.08) at $40/MWh and the respective IT load. These are the energy savings alone — the minimum recoverable cost confirmed by public utility and ESG data. The full EA value includes what those recovered megawatts could do instead: run AI workloads. At $38M per recovered MW-year (Synestra's full-stack EA value intensity), a 500 MW campus at this PUE gap carries $1.52B in total EA opportunity annually. The energy floor is $14M. The gap between the two is compute capacity that overhead is consuming instead of generating revenue. Review the commercial thesis.
The competitive frame: An operator running at EA Score 73 can serve the same AI compute demand as one running at 55 — at lower infrastructure cost and higher margin. As AI compute capacity expands, operators with higher EA Scores gain a pricing advantage that compounds across every tenant negotiation. A score of 55 is not a neutral starting point. It is a cost structure that a higher-scoring competitor can undercut.
Your Campus
Adjust the inputs below using your facility's published or internal specifications. The calculation uses the same methodology as the operator profiles above.
This is a scenario model based on the inputs above. It does not claim guaranteed outcomes, universal EA improvement, or Synestra customer operating data. Actual recoverable value depends on facility design, workload mix, controls maturity, and operating model. Full methodology.
EA Score Methodology
The Economic Availability Score is a 0–100 index measuring the percentage of designed economic capacity an AI infrastructure campus is actually delivering. Synestra publishes the methodology. The score is calculated consistently across operators using the same formula.
Facility Score (public data)
Measures the percentage of facility power that reaches IT equipment. A PUE of 1.28 means 78% of power reaches compute — a Facility Score of 78. Calculable from published ESG data for any operator that discloses PUE.
Full EA Score (requires deployment)
Adds the compute yield dimension — what fraction of IT power is actually delivering useful workload output. Industry average AI factory GPU utilization is approximately 70%, producing a Full EA Score of roughly 55 for a campus with a Facility Score of 78.
Synestra Target
What is achievable with PUE improvement to the operator's stated target and workload coordination raising GPU utilization toward 88%. The gap between the current Full EA Score and the Synestra Target is the recoverable EA opportunity.
What the score means
The PUE gap captures the landlord-side recovery opportunity — power consumed by mechanical systems that could be recovered and redeployed as additional IT load or cost reduction. It does not capture BTM dispatch optimization, compute yield improvement, or workload scheduling efficiency, which are additional EA recovery levers. The full EA opportunity at a given campus is larger than the PUE gap alone.