ESG & Sustainability
The fastest path to better ESG
is operational efficiency.
Data centers are under pressure on energy, water, and carbon from every direction — tenants, investors, regulators. The industry's response is mostly new hardware. But the largest source of waste in AI infrastructure is not the hardware. It is the gap between what the hardware could do and what it actually does. That gap is the ESG problem. Closing it is also the financial problem. Synestria closes both at once.
Typical operational waste
25–40%
Of IT energy draw at a large AI campus produces no billable compute output, due to operational gaps across power, cooling, network, and scheduling.
CO₂ recovery at 200 MW
100K+ tonnes
Annual CO₂e reduction achievable through EA improvement alone on a 200 MW campus — without adding a single renewable energy contract.
ESG reporting impact
Same fix
The EA recovery event that returns lost revenue to the operator simultaneously reduces energy and carbon intensity reported under GRI, SASB, and CDP frameworks.
How ESG is measured in data centers
The metrics investors and tenants watch.
ESG reporting for data center operators pulls from a combination of industry efficiency standards and third-party frameworks. These are the numbers your tenants ask for, your investors benchmark, and regulators are beginning to require.
Energy
Power Usage Effectiveness
PUE
The ratio of total facility power to IT equipment power. A PUE of 1.0 is theoretically perfect. Best-in-class campuses reach 1.1–1.2. Industry average is 1.4–1.6. PUE measures cooling and power delivery efficiency. It says nothing about whether the IT load is doing useful work.
PUE = Total Facility Power ÷ IT Equipment Power
Water
Water Usage Effectiveness
WUE
Annual water consumption divided by IT energy consumed. WUE is a defined ESG reporting metric, but its applicability varies significantly by cooling architecture — evaporative and hybrid-cooling facilities track WUE actively; modern closed-loop liquid and air-cooled hyperscale campuses report minimal water draw under this metric.
WUE = Annual Water Usage (L) ÷ IT Energy (kWh)
Carbon
Carbon Usage Effectiveness
CUE
Total carbon emissions from facility operations divided by IT energy. Includes Scope 1 (on-site generators), Scope 2 (purchased electricity), and increasingly Scope 3 (supply chain). Lower PUE improves CUE, but CUE also tracks the carbon cost of every wasted IT compute hour at the grid carbon intensity rate.
CUE = Total CO₂ Emissions ÷ IT Energy (kWh)
Reporting
GRI Standards
Global Reporting Initiative
The most widely used ESG reporting framework globally. GRI 302 (Energy) and GRI 303 (Water) are directly applicable to data center operators. Tenant customers with public ESG commitments increasingly require GRI-aligned disclosure from their data center providers as part of supplier due diligence.
Reporting
SASB Standards
Technology & Communications Sector
SASB's Technology & Communications standard includes data center-specific metrics: energy consumed, percentage from renewables, and water consumed in water-stressed regions. Public companies and their major suppliers are increasingly expected to report against SASB standards.
Reporting
CDP & TCFD
Carbon Disclosure / Climate Financial Risk
CDP (formerly Carbon Disclosure Project) collects self-reported emissions data annually. TCFD (Task Force on Climate-related Financial Disclosures) is now embedded in regulatory requirements in the EU, UK, and increasingly in US SEC climate disclosure rules. Both are creating mandatory reporting pressure throughout the data center supply chain.
Impact at scale — 200 MW campus
What EA improvement actually moves on the ESG scorecard.
These are modeled estimates based on industry-standard figures: US grid carbon intensity of 0.386 kg CO₂e/kWh (EPA eGRID 2023), and a typical operational efficiency gap of 30%. A 200 MW campus is a mid-to-large AI facility — comparable to many hyperscale deployments currently under construction. Actual impact varies by site, power mix, and operational baseline.
200 MW AI Campus — Annual ESG Impact
Baseline vs. EA-optimized operations
Total energy draw
1.752 TWh/yr
1.226 TWh/yr
526 GWh saved
Productive compute output
~70% of IT load
~92% of IT load
+22 percentage points
Carbon emissions (Scope 2)
677,000 tonnes CO₂e
473,000 tonnes CO₂e
204,000 tonnes eliminated
PUE equivalent improvement
1.50 (industry avg)
Unchanged
PUE doesn't capture this
Carbon offset equivalent
—
—
~44,000 cars off road/yr
PUE measures how efficiently you deliver power to the IT load. It does not measure whether the IT load is doing anything with it. A campus with a PUE of 1.15 and 35% operational waste has a worse sustainability profile than one with a PUE of 1.40 and 10% waste. No ESG framework currently captures the difference. Synestria does.
The sustainability gap is not in the cooling towers. It is in the operational layer between the hardware and the workload — the gap that EA is designed to close.
Why this is different
Sustainability as a financial return, not a cost.
Operators are used to treating ESG improvements as investments — renewable energy contracts, new cooling infrastructure, water recycling systems. EA improvement is the exception. Every megawatt hour of waste eliminated is simultaneously an ESG improvement and a financial recovery. The sustainability win and the revenue win are the same event.
🌱
ESG improvement
Every EA recovery event reduces the energy and carbon associated with a unit of compute output. That shows up directly in your GRI 302 energy intensity number, your CUE, and your CDP disclosure — without adding renewable capacity or installing new cooling equipment.
Lower Scope 2 emissions per GPU hour
Improved carbon intensity per unit of compute
Improved CUE for tenant disclosures
Verifiable, data-backed ESG reporting
Supports SBTi alignment without additional capex
📈
Financial recovery
The wasted energy you are already paying for is also the revenue you are not billing. When EA improvement recovers operational efficiency, it recovers both simultaneously. No new spending required. The gain-share model means Synestria only earns when you do.
Recovered compute capacity = recovered billable hours
Reduced energy cost per unit of output
Lower cooling energy spend per job completed
Improved tenant SLAs without additional infrastructure
Gain-share billing: Synestria earns from the improvement
How EA improvement creates the ESG outcome
A single operational event. Two measurable results.
The consequence chain below shows how a typical EA recovery event maps to an ESG improvement. Synestria detects the degradation before it becomes waste, intervenes in real time, and the improvement flows through both the financial and ESG ledger simultaneously.
Consequence chain: thermal anomaly to ESG recovery
One event, traced from detection to dual outcome. This plays out dozens of times per day across a large AI campus.
01
Cooling circuit begins to degrade
A cooling unit loses 8% capacity due to a fouled heat exchanger. The EA Consequence Graph detects the thermal drift 23 minutes before it reaches the alarm threshold on the existing BMS. The circuit is still reporting healthy.
02
Without intervention: waste begins
Without early detection, the cluster enters thermal throttle. GPU compute drops to 70% performance. The job runtime extends by 40%. Cooling systems compensate by running at overcapacity — drawing additional power and water. The facility uses more energy to produce less useful output. PUE stays the same. The ESG cost is invisible.
03
With Synestria: early intervention
The consequence model issues a recommendation to reroute cooling load before the throttle fires. The affected circuit is isolated. The GPU cluster maintains full throughput. The job completes on schedule. Cooling systems run at productive baseline — not compensating for a problem that was not caught in time.
04
Dual outcome: ESG + financial
The energy that would have been consumed by overcooling is not consumed. The GPU hours that would have been lost to throttle are billed. The ESG report reflects lower energy intensity per compute hour. The operator recovers the revenue from compute hours that would have been lost. Synestria earns its gain-share. Everybody wins.
ESG reporting alignment
Where EA improvement shows up in your disclosures.
Synestria does not produce ESG reports. But the operational data it captures — energy by system, by domain, by event, with timestamps and consequence attribution — is exactly the data that supports accurate ESG disclosure.
📊
GRI 302 — Energy
EA improvement reduces energy consumed per unit of compute output. Synestria's telemetry provides domain-level energy attribution — power, cooling, compute — with the resolution needed to support GRI 302-3 (energy intensity) reporting accurately.
💧
GRI 303 — Water
GRI 303 applies primarily to facilities with significant water draw. For campuses using closed-loop cooling architectures — the standard for new hyperscale builds — Synestria's thermal domain data provides the operational context needed to support accurate GRI 303 disclosure, including documentation of cooling system type and actual water consumption.
🌍
Scope 2 Carbon (GHG Protocol)
Energy reduction from EA improvement flows directly to Scope 2 emissions. Synestria's per-domain, per-event energy attribution supports location-based and market-based Scope 2 accounting with time-stamped, auditable data.
📋
SASB — Technology & Communications
SASB TC-SI-130a.1 (energy consumed) and TC-SI-140a.1 (water consumed) are the primary data center metrics. EA operational data maps directly to both. Percent of renewables is unchanged by EA improvement — but the denominator (total energy consumed) decreases, improving the ratio.
🔭
CDP Disclosure
CDP's climate questionnaire requires data on energy use, emissions, and management approaches. The operational transparency that Synestria provides — consequence-attributed energy events, not just totals — supports higher-quality CDP responses and demonstrates active emissions management beyond PUE tracking.
🎯
Science Based Targets (SBTi)
SBTi requires verifiable, year-over-year emissions reductions. EA-driven operational efficiency improvements produce reductions that are attributable, auditable, and not dependent on purchased offsets or renewable energy credits. They are real reductions in energy consumed per unit of economic output.
Start with your EA baseline.
We establish your Economic Availability baseline in 30 days. That baseline quantifies your operational efficiency gap — and the ESG improvement and revenue recovery available when you close it. No obligation if we find no material gap.
