Hypervisor-Native Thermal & Cooling Intelligence
Compute density is only one aspect of modern datacenter infrastructure; another is maintaining that density effectively, safely, and sustainably.
Cooling, sensors, and sustainability metrics are treated as afterthoughts by traditional provisioning platforms, which rely on external systems.
By integrating thermal and cooling intelligence directly into the hypervisor and integrating it into the same zero-touch provisioning workflow that distributes workloads, controls firmware, and maximizes power, Karios AtlasFlow alters that paradigm.
Redefining Power Control for Next-Generation Infrastructure
The first solution to provide real-time visibility of rack-level cooling loops and immersion cooling tanks inside the hypervisor is Karios AtlasFlow.
With AtlasFlow, operators can view thermal load, coolant flow, and tank performance data alongside virtual machine (VM) and container workloads, eliminating the need for separate Building Management Systems (BMS) or external dashboards.
This unified view bridges the gap between compute and thermal management, enabling smarter scheduling, workload placement, and proactive failure prevention.
The first solution to provide real-time visibility of rack-level cooling loops and immersion cooling tanks inside the hypervisor is Karios AtlasFlow.
With AtlasFlow, operators can view thermal load, coolant flow, and tank performance data alongside virtual machine (VM) and container workloads, eliminating the need for separate Building Management Systems (BMS) or external dashboards.
Through open standards and integrations, Karios AtlasFlow directly feeds into Karios Atlas, our unified infrastructure intelligence layer. These include:
- Modular Industrial IoT standards for sensor onboarding, or PICMG
- Redfish: RESTful telemetry and out-of-band management
- Open-source baseboard management, or OpenBMC
- Intel vPro: Deep remote management integration for non-BMC devices
- PiKVM: afordable KVM-over-IP onboarding for small-form-factor and edge computing
AtlasFlow isn’t just monitoring it’s a sensor and IoT ingestion engine integrated into the hypervisor.
Data streams are visualized instantly in the Atlas UI or accessed via API, unifying compute, thermal, and infrastructure telemetry in one pane of glass.
Karios AtlasFlow is in line with important environmental scoring metrics and was developed with sustainability mandates in mind:
- PUE (Power Usage Effectiveness): Ongoing insights to lower overall power consumption
- CUE (Carbon Usage Effectiveness) – Tie cooling efficiency directly to CO₂ impact
- Water Usage Effectiveness (WUE): Immersion cooling tracking and optimization in real time
By directly integrating with Karios’ power-management integration, AtlasFlow ensures efficiency and compliance by automatically modifying workloads and thermal management based on real-time data.
Cooling failure is a risk to business continuity rather than just a thermal problem.
In order to identify failure precursors before they occur, AtlasFlow incorporates AI-driven predictive monitoring directly at the hypervisor level, examining vibration patterns, pump cavitation, and bearing wear.
Karios’ zero-touch provisioning workflows incorporate this predictive capability to guarantee early alerts and continuous uptime.
Why It Matters
In traditional datacenters, IT and facilities work in silos IT manages workloads while facilities handle cooling.
Karios AtlasFlow eliminates that divide by embedding thermal and sustainability intelligence directly into the hypervisor so that:
- 01 Workloads are provisioned with full thermal awareness
- 02 Sustainability metrics are continuously tracked and auditable
- 03 Cooling issues are predicted before they cause downtime
- 04 Compliance is achieved without third-party overhead
AtlasFlow redefines Zero-Touch Provisioning — not only for se
Karios vs. Legacy Platforms
- Karios Forge
- VMware
- Nutanix
- Proxmox
Capability | Karios Forge | VMware | Nutanix | Proxmox |
| Hypervisor-Level Cooling Visibility | ✅ Native immersion & rack-cooling visibility | ❌ Not supported | ❌ Not supported | ❌ Not supported |
| BYOS – Bring Your Own Sensors | ✅ Fully agnostic, IoT/industrial mapping | ❌ Proprietary lock-in | ❌ OEM-limited | ❌ Not supported |
| Integration with Open Standards (PICMG, Redfish, OpenBMC, vPro, DASH, PiKVM) | ✅ Full integration & telemetry unification | ❌ Partial Redfish only | ❌ Proprietary APIs | ❌ Not supported |
| Unified Infrastructure Telemetry (Compute + Thermal) | ✅ Single-pane Atlas UI/API | ❌ Requires external tools | ❌ Third-party stack | ❌ Not supported |
| Sustainability Metrics (PUE, CUE, WUE) | ✅ Native scoring, compliance-ready | ❌ Add-ons required | ❌ Add-ons required | ❌ Not supported |
| Predictive Cooling Failure Analytics (AI-driven) | ✅ Built-in hypervisor analytics | ❌ Not supported | ❌ Not supported | ❌ Not supported |
| Zero-Touch Provisioning with Thermal Context | ✅ Unified workload, firmware, cooling workflows | ❌ Compute-only | ❌ Limited | ❌ Basic |
| Non-BMC Device Support (Intel vPro, DASH, PiKVM) | ✅ Edge/laptop provisioning support | ❌ Not supported | ❌ Not supported | ❌ Not supported |
| Compliance & Audit Readiness (Energy + Sustainability) | ✅ Built-in scoring system | ❌ Not supported | ❌ Not supported | ❌ Not supported |
Key Takeaways
- VMware – a leader in virtualization, but it relies on costly overlays and lacks native cooling, sustainability, or IoT integration.
- Nutanix – Excellent HCI, but no consideration for sustainability or heat.
- Proxmox – Lightweight and open, yet compute-only.
- Karios AtlasFlow – The only hypervisor that combines predictive analytics, IoT onboarding, sustainability metrics, and thermal intelligence into a single, zero-touch provisioning framework.