Windows
LibreHardwareMonitor sensors plus per-process attribution. Workstations and Windows Server included.
Product · ESG agent
From source telemetry to defensible Scope 2.
Most technology ESG reporting breaks at one of two points: data quality, or access to it. The Horizon Portal ESG agent is a cross-platform telemetry layer that turns infrastructure activity into evidence-ready energy, Scope 2, and cost figures — with the chain of custody every link of the way.
One agent, the whole heterogeneous fleet.
A single ESG agent runs across the platforms most operators actually have on the floor. Numbers come from primary sources where the platform exposes them, calibrated estimates where it doesn't — and the method label always travels with the figure.
macOS
powermetrics on Apple Silicon (SoC + integrated GPU); Intel Macs use HWMON-equivalent counters.
Linux
IPMI for whole-system draw, RAPL for CPU+DRAM, NVIDIA / AMD GPU counters where present, lm-sensors as fallback.
Proxmox
Host-level power from the underlying Linux stack, with per-VM allocation by CPU usage. The allocation method is logged with the figure.
pfSense
Edge device draw, so the perimeter appliance shows up in the same fleet view as the servers behind it.
In planning
Other hypervisors (VMware ESXi, Hyper-V) and additional operating systems are on the roadmap. Tell us what you need next on the contact form.
A 7-day measured run, on our own infrastructure.
These figures come from our own kit, not a customer reference. We run the ESG agent on our own infrastructure first — partly because the only credible answer to "show me the numbers" is to have your own ones to show, partly because we wouldn't ship anything else.
80.1 kWh
Measured across 10 endpoints, 7-day rolling window.
16.6 kg CO₂e
UK grid factor; full provenance in the disclosed-factors table.
£24.03
Ofgem UK 2024 unit rate; swap in your contracted rate at deployment.
10
Mixed: Windows, macOS, Linux, Proxmox, pfSense.
Same shape of output you receive from the ESG Evidence Pilot, on your own fleet.
Measurable, methodology-transparent, practical.
The frame is straightforward: every link in the chain has a method, every method has an accuracy band, both get reported alongside the number. Vendor averages don't survive a hostile auditor; primary data with disclosed methods does.
- Capture. Per-endpoint kilowatt readings from primary sources where available (IPMI, RAPL, GPU counters, powermetrics), calibrated estimation where they aren't. Method labelled per endpoint.
- Allocate. Host-level draw split across VMs by CPU usage on Proxmox; per-process attribution on workstation OSes. Allocation method logged with each figure.
- Convert. kWh → kg CO₂e using the grid-emission factor for your jurisdiction and reporting period (or your contracted PPA / green-tariff factor if you have one). kWh → £ using your contracted electricity rate.
- Aggregate. Per-endpoint, per-fleet, per-time-window. Daily, 7-day, 30-day, and annualised projections — all clearly labelled as projections, not actuals.
- Export. CSV with timestamps, workload IDs, factor sources, and method-accuracy badges. The trail an auditor follows.
The full method-by-method breakdown — what IPMI measures versus RAPL versus calibrated estimation, and the accuracy band on each — sits on the Measurable AI page.
What the agent produces
Live fleet telemetry
Live wattage per endpoint, with a CPU / RAM / GPU breakdown. Method-accuracy badges on every reading: green for primary counters, amber for mixed methods, red for estimation.
Period rollups
Fleet kWh, kg CO₂e, and £ on 24-hour, 7-day, and 30-day windows, with annual projections. Same numbers across the dashboard, the API, and the exported report.
Auditor-ready ESG report
Per-endpoint detail, GPU detail, VM allocation method, and methodology footer. The kind of artefact a sustainability team can hand to assurance without an apology.
Period comparison
Week-over-week and month-over-month deltas for kWh, kg CO₂e, and £. Useful when something changes — a hardware refresh, a workload shift, a tariff renegotiation — and you want to see it in the numbers.
Cost allocation
Across companies, departments, or customers, based on actual measured consumption rather than headcount or floor area. Useful for MSPs reporting per-customer impact.
CSV export
One row per endpoint per measurement window. Timestamps, workload IDs, factor source, unit-cost source. The substrate that reasonable assurance can rest on.
A pilot, then per-agent licensing.
ESG Evidence Pilot
£3,000 ex VAT
Fixed fee. Up to 10 endpoints, 1–2 weeks, 7-day measured baseline + auditor-ready CSV + readout.
See the ESG Evidence PilotContinuing telemetry
£1 per agent / month
Billed only on agents you choose to keep running after the pilot. Stop at the end of the 7-day baseline and keep the report; or keep the agents live and the numbers refreshing.
Horizon Portal itself is self-hosted on your infrastructure; the ESG agent runs alongside, inside the same perimeter.
What this is, and isn't
- Measurable is not low-carbon. The agent gives you the kilowatt number; whether it's acceptable is a decision the figures can inform but not make.
- Scope 2 only, today. Purchased electricity for IT infrastructure. Scope 1 (direct fuel, refrigerants) and Scope 3 (everything else) are out of scope.
- Estimation has a place. Not every endpoint exposes a primary counter. When calibrated estimation is the best signal available, we use it, and we flag it. A clearly-labelled estimate is more defensible than an unlabelled measurement.
- Audit-grade is not assurance-grade. SECR and CSRD require reasonable-assurance disclosures, not a stamp from us. Primary data is the substrate that reasonable assurance can rest on; the judgement still belongs to the auditor.
A 30-minute ESG evidence overview.
No slides. We walk through the dogfood baseline, the methodology, what the agent looks like running on a mixed fleet, and what a pilot scoped to your environment would cost.