Built for the systems you can’t afford to get wrong.
AquaWatt DC OS sits next to mission-critical infrastructure. It is designed from the ground up around data protection, strong access control, network isolation, and governed, auditable actions.
Data protection
- Encryption in transit
- Encryption at rest
- Tenant isolation
- Secrets management
- Key rotation
- Backup & disaster recovery
Access & identity
- Least-privilege access
- Multi-factor authentication
- SSO via SAML / OIDC
- Role-based access control
- Device identity
Network & operations
- Network segmentation
- Private connectivity
- Secure edge-to-cloud comms
- Signed software updates
- Immutable audit logs
- Security-event export
- Vulnerability management
Intelligence in the cloud, control on the ground.
AquaWatt separates the analytics plane from the control plane. Telemetry flows outbound from the edge to power forecasting and optimization, but changes to equipment are gated by local policy validation and operator approval. Safety-critical logic runs at the edge and continues to protect equipment even if the cloud connection is interrupted.
Compliance by design.
The platform is being engineered to support the frameworks enterprise data center operators care about. Our architecture, controls, and processes are aligned to these standards as we mature.
SOC 2
Security, availability, and confidentiality controls.
Designed to support
ISO 27001
Information security management systems.
Designed to support
ISO 50001
Energy management systems.
Designed to support
Important: AquaWatt is an early-stage company. AquaWatt DC OS is designed to support enterprise security and compliance programs, but the platform is not yet certified for SOC 2, ISO 27001, or ISO 50001. These frameworks describe the standards we are building toward, not certifications we currently hold.
Have a security or compliance question?
We’re happy to walk your security and infrastructure teams through our architecture and roadmap.
Optimization is the last priority — by design
An efficiency system that can cause an outage is not an efficiency system. AquaWatt DC OS resolves every conflict against a fixed priority order, and resource savings sit at the bottom of it.
Safety hierarchy
When two objectives disagree, the higher priority always wins.
- 1Human safety
- 2Equipment safety
- 3Uptime and redundancy
- 4Workload SLA compliance
- 5Regulatory compliance
- 6Resource optimization
Hard safety boundaries
AI models cannot override configured hard constraints. A constraint is a wall, not a preference the optimizer may weigh against savings.
Fail-safe behaviour
If data quality, connectivity or model confidence drops below approved limits, affected automation pauses, a safe state is maintained or restored, authorized personnel are alerted and diagnostic information is preserved.
Independent policy engine
Safety-policy evaluation is logically separated from the models that generate optimizations, so the system that proposes an action is never the system that approves it.
Change-rate limits
How quickly an eligible setpoint may move is bounded. Gradual change keeps thermal and electrical systems inside the envelopes operators already trust.
Control conflict prevention
Conflicting commands from operators, automation systems or existing controllers are detected rather than silently applied on top of one another.
Shadow mode
New models and policies run alongside live operation with no control privileges, so their judgement can be audited against reality before it counts.
Canary deployment
Automation is enabled first for limited equipment, zones or time windows, then widened only once behaviour is proven at the smaller scale.
Describes the safety architecture AquaWatt DC OS is being designed around. Autonomous control is excluded from the first release; early deployments operate in advisory and human-approval modes.
Who can act, and what happens when something goes wrong
A platform that can influence cooling, water and power has to be explicit about permission and accountability. AquaWatt DC OS is designed so that authority is granular, actions are attributable, and a critical event becomes a managed incident rather than an unread alert.
Enterprise authentication
Enterprise authentication with multi-factor authentication is a foundational requirement, not an upgrade tier.
Single sign-on
SAML or OIDC-based single sign-on so accounts live in the identity system your security team already governs.
Attribute-based restrictions
Beyond roles, permissions can be narrowed by facility, system, region, data type and the risk level of the action itself.
Session and access audit
Authentication events and privileged activity are logged, so who did what — and who could have — is answerable after the fact.
Role-based access, built around how infrastructure teams are actually organised
An executive reviewing a portfolio, an engineer approving a cooling change and an auditor reading history need very different powers. The initial role set reflects that.
Alerts reach people where they already work
Operations teams do not live in one tool. Notifications are deliverable across the channels a facility already runs on.
Escalation policies
Escalation is configured by severity, facility, system and time of day, so a 3 a.m. thermal excursion does not follow the same path as a routine advisory.
Incident creation
Critical alerts do not stop at a notification. They open an operational incident that has an owner and a state.
Incident timeline
Every relevant alert, control action, telemetry change and communication is assembled into a single ordered timeline of what actually happened.
Post-incident analysis
The platform drafts an incident analysis with recommended corrective actions, so the review starts from evidence rather than from memory.
Describes the planned access control, notification and incident management design of AquaWatt DC OS. Individual capabilities are delivered across releases and are not all available in the initial pilot build. AquaWatt is not currently certified under SOC 2, ISO 27001 or ISO 50001.