The AI-Native DCIM That Runs Anywhere
Modern asset, capacity and change management with agentic AI on top - multi-tenant SaaS in a week, or fully sovereign inside your perimeter with the AI still included. SaaS, dedicated private cloud, or air-gapped on-premises - same platform, your boundary.
The Physical Estate Is the Last Unautomated Risk
Legacy DCIM can't deliver AI to regulated buyers
Most incumbent platforms like Nlyte and Sunbird were designed for cloud-connected deployments, creating real friction for APRA and RBNZ-regulated entities that require data sovereignty. Feeding asset telemetry and topology data to a foreign cloud model may create material compliance risk and requires significant risk and legal assessment to satisfy APRA and RBNZ obligations. Many CISOs at regulated entities will raise data sovereignty concerns before a pilot begins.
Spreadsheets break under AI-era rack densities
Modern GPU and high-density compute deployments make power, cooling, and space capacity difficult to track in Excel. In high-density environments, capacity models can become unreliable within weeks of a major deployment without automated updates, blind spots compound, and by the time a capacity constraint surfaces, the change is already approved. The estate your regulator scrutinizes most may run on the least reliable data.
Audit evidence is still assembled by hand
APRA CPS 230 and CPS 234 and RBNZ outsourcing rules increasingly expect continuous control evidence — not a binder assembled in the fortnight before an audit. DC operations teams face growing manual evidence burdens even as headcount pressures mount.
One Platform. Every Deployment Model. AI Inside Your Boundary.
CenterOS brings AI inference, models, and the data they learn from entirely inside the customer's control perimeter, built on a complete and mature data model of assets, power, cooling, and physical connections that makes capacity forecasting, anomaly detection, and impact analysis credible rather than decorative. It delivers AI-assisted change tracking, a continuous evidence trail an auditor will accept, and agentic AI that carries routine operational load under human approval. One platform, three deployment models: multi-tenant SaaS, dedicated private cloud, and fully sovereign on-premises including air-gapped.
Sovereign AI that runs where your policy demands
CenterOS runs inference, models, and the data they learn from entirely inside your control perimeter — including fully air-gapped on-premises deployments on your own GPUs. Agentic AI is not a cloud-attached add-on; it is included at every deployment tier. Regulated operators can use the AI features they're paying for without compromising their data sovereignty requirements.
A complete data foundation that makes AI credible
Six-dimensional capacity tracking across power, cooling, space, weight, connectivity, and compute (plus custom dimensions) — backed by a VictoriaMetrics monitoring datalake speaking SNMP, Modbus, BACnet, MQTT, IPMI, and Prometheus. The AI works from a structured, comprehensive model of your estate, which is what makes capacity forecasting and anomaly detection useful rather than decorative.
Continuous audit evidence generated by the system
Every AI recommendation, every change workflow, and every asset state carries an immutable audit trail. CenterOS produces continuous control evidence designed to support APRA CPS 230, CPS 234, and RBNZ outsourcing obligations — generated automatically by the platform, not assembled under pressure before an audit. Your compliance team gets evidence on demand.
From Deployment to Audit-Ready in Days, Not Months
Deploy in the model that fits your policy
Choose multi-tenant SaaS and be productive in a week, spin up a dedicated private cloud with contractual SLAs, or deploy fully on-premises — air-gapped if required. Your deployment model does not limit your AI capability.
Ingest your estate and existing data
CenterOS connects to your monitoring infrastructure via SNMP, Modbus, BACnet, MQTT, IPMI, and Prometheus. Legacy migration tooling moves your existing DCIM data, spreadsheets, and asset records into a single authoritative model without a rip-and-replace project.
Let CenterOS AI operate under human approval
Ask your data center questions in plain English. CenterOS AI plans capacity, correlates incidents, and drafts change workflows — all submitted for human approval before execution. Every recommendation is traceable, explainable, and ready for an auditor the moment it is acted on.
Three Deployment Models. One Capability Set.
SaaS
- Multi-tenant cloud deployment — productive in under a week
- Full asset lifecycle, capacity tracking, and monitoring datalake
- CenterOS AI included — agentic capacity planning and incident correlation
- Continuous audit trail and change management
- Ideal for: colocation operators, MSPs, and enterprise teams modernizing off spreadsheets
Private Cloud
- Dedicated single-tenant deployment with contractual SLAs
- All SaaS capabilities plus enhanced data isolation
- CenterOS AI on dedicated infrastructure — data does not share compute with other tenants
- Ideal for: regulated operators requiring stronger isolation without full on-premises commitment
Sovereign On-Premises
- Fully on-premises, including air-gapped deployment on your own hardware
- AI inference and models run on your own GPUs inside your control boundary
- All platform capabilities — no features withheld for on-premises deployment
- Ideal for: APRA- and RBNZ-regulated entities, government agencies, and defence-adjacent operators
Frequently Asked Questions
See CenterOS Running Inside Your Boundary
Join our paid proof-of-concept program and have a production-grade, AI-native DCIM environment running on your infrastructure — with your data — in under two weeks.