Skip to main content
Interactive platform simulation

The Resource Operating System
for AI Data Centers

AquaWatt DC OS uses artificial intelligence, digital twins and safety-constrained optimization to coordinate power, water, cooling and compute—helping data centers increase resource efficiency without compromising uptime.

Platform simulation

Facility Power

24.7 MW

Cooling Load

8.2 MW

Water Flow

380 L/min

PUE

1.38

WUE

0.82

GPU Utilization

78%

Thermal Risk

Low

Carbon Intensity

142 gCO₂/kWh

AI compute is colliding with
physical infrastructure.

POWER

AI and high-density GPU workloads create increasing electrical demand.

WATER

Cooling strategies can consume significant water and become especially sensitive in water-constrained regions.

THERMAL

Higher rack density makes thermal conditions an operational and business constraint.

FRAGMENTATION

BMS, DCIM, workload schedulers, utility signals and sustainability systems rarely make decisions together.

Power, water and compute are no longer separate problems.

One intelligent system.
Six coordinated stages.

Nine integrated modules.
One operating system.

Resource Command Center

Live visibility across power, water, cooling, carbon, cost and workload state.

Data Integration Fabric

Connect BMS, DCIM, SCADA, meters, sensors, workloads, weather and utility data.

Forecasting + Anomaly Intelligence

Forecast resource demand and detect abnormal operating conditions.

Data-Center Digital Twin

Simulate operating strategies before making changes.

Multi-Objective Optimization Engine

Balance energy, water, carbon, cost, peak demand, hardware wear, workload delay, reliability.

Workload Resource Orchestrator

Coordinate eligible workloads with facility and grid conditions.

Autonomous Control + Safety Layer

Policies, approvals, rollback, emergency stop and customer-controlled autonomy.

Sustainability + Financial Intelligence

PUE, WUE, CUE, ERE, savings, cost and environmental reporting.

AquaWatt AI Copilot

Natural-language investigation, scenario analysis and operational planning.

Powered by

AquaWatt DC OS

Explore all nine modules
AI Copilot

An operator’s copilot that explains before it acts.

Ask questions in plain language. The copilot forecasts, simulates against a digital twin, and returns recommendations with the reasoning and trade-offs behind them — always waiting for your approval before anything changes on the floor.

  • Natural-language questions about your facility
  • Recommendations backed by simulation, not guesswork
  • Every action reversible and fully auditable
See how it works
AquaWatt Copilot
Illustrative demo

Cooling load in Hall 3 is climbing. What can we do without risking thermal limits?

I forecast Hall 3 inlet temps approaching your soft limit in ~40 min. Shifting 6% of non-urgent batch jobs to Hall 1 and raising chilled-water setpoint by 0.5°C keeps you within limits and reduces cooling energy.

Projected energylower cooling draw
Reliabilitywithin safe margins
Confidencesimulated on digital twin

Show me the trade-offs before I approve.

Here is the side-by-side: option A prioritizes energy, option B prioritizes thermal headroom. Both preserve uptime. Nothing changes until you approve, and I can roll back instantly.

Safety architecture

Autonomy you dial in, safety you never give up.

You choose how much authority to delegate — from pure monitoring to supervised autonomy — and can change it at any time. Underneath every level, the same safety guarantees always hold.

L0

Monitor

AquaWatt observes and reports. No actions are taken.

L1

Advise

The platform recommends actions with full reasoning; operators decide.

L2

Assisted

Operators approve recommendations, which AquaWatt then applies and verifies.

L3

Supervised autonomy

Within operator-defined guardrails, routine actions run automatically, always reversible.

Always-on guarantees

Hardware interlocks and equipment safety limits are always respected.
Safety-critical logic runs locally at the edge and keeps working if the cloud is unreachable.
Every change is simulated on a digital twin before it is proposed.
Operators can override or roll back any action instantly.

More than a dashboard. An operating system.

Traditional DCIM and building-management tools tell you what happened. AquaWatt DC OS forecasts what’s next, recommends what to do, and — with your approval — acts and verifies the result.

Dimension
Traditional DCIM / BMS
AquaWatt DC OS
Scope of optimization
Isolated systems (power, cooling, IT managed separately)
Whole-system optimization across power, cooling, water, and compute
Intelligence
Dashboards and thresholds; humans interpret everything
AI forecasting, digital twins, and multi-objective optimization
Explainability
Alerts without reasoning
Every recommendation explained with expected impact and trade-offs
Action & control
Manual changes, limited guardrails
Governed autonomy with approvals, rollback, and audit trail
Sustainability
Reported after the fact, often modeled
A live operating variable with verified, traceable outcomes
Water
Rarely modeled or optimized
Monitored and weighted by local water stress

Initial Pilot Targets

Targets, not guarantees. Results depend on facility conditions.

Cooling-energy reduction
0–0%
Total facility-energy reduction
0–0%
Potable-water reduction
0–0%
Peak-demand reduction
0–0%
GPU-utilization improvement
0–0%
Thermal-incident reduction
0–0%
Reporting-time reduction
0–0%
Safety Targets

0

optimization-related SLA violations

0

critical unsafe control actions

Results will depend on facility design, operating conditions, workload profile, equipment, climate, water strategy and existing efficiency.

Calculate the opportunity.

Facility Parameters

20 MW
$0.08
$5.0M
$500K
1.50

Illustrative estimate — not a guaranteed AquaWatt result

Illustrative annual energy savings opportunity

$1.4M

Illustrative water savings opportunity

$113K

Estimated annual operating-cost opportunity

$2.7M

Request a Facility Assessment

Prove AquaWatt in a
live data center.

Month 1

Discovery + Integration

Month 2

Data Validation + Baseline

Month 3

Forecasting + Anomaly Intelligence

Month 4–5

Optimization Recommendations

Month 6

Savings Verification + Expansion