A secure building data platform for CIOs, IT and data teams
A secure building data platform brings meters, sensors and building systems under one governed layer IT can approve. Aqueduct ingests BACnet, Modbus, MQTT, M-Bus, pulse and LoRaWAN data into aQ Core, runs in the AWS cloud or on-premises with on-prem AI on NVIDIA infrastructure, and is backed by ISO/IEC 27001, ISO 9001:2015 and SOC 2 Type 2.
One governed layer for building data, deployed the way your policy requires.
A posture your review board recognizes
- Certifications
ISO/IEC 27001, ISO 9001:2015 and SOC 2 Type 2 behind the platform and the team.
- Controlled actions
Equipment commands only where control is enabled, with owner approval and full command history.
- Data quality shown
Device health and coverage are visible, so nobody reports on data that is not there.
Cloud or on-premises, with AI where your data lives
- AWS cloud
A managed cloud deployment on AWS, from an AWS Advanced Tier Partner.
- On-premises
Aqueduct on your infrastructure, with on-prem AI on NVIDIA infrastructure where required.
- Network paths
LoRaWAN gateways and IP devices backhaul over Ethernet, Wi-Fi or LTE, agreed with IT first.
Open protocols in, enterprise systems out
- Device agnostic ingest
BACnet, Modbus, MQTT, M-Bus, pulse and LoRaWAN, from existing or newly specified devices.
- One data model
Sites, floors, zones and devices in aQ Core, so every module reads the same records.
- CMMS and beyond
Tickets route into your CMMS, and data integrates with EAM, IWMS, historians and BI.
Building data for IT, before and after Aqueduct
- Building systems appear on the network before IT knows.
- Every system brings its own cloud and its own login.
- Building data is locked in proprietary historians.
- Regulated sites cannot send data off premises.
- Equipment overrides leave no audit trail.
The Aqueduct way- One governed platform, reviewed by IT before rollout.
- One data model in aQ Core across every system.
- Data integrates with CMMS, EAM, IWMS and BI.
- On-prem deployment, with on-prem AI on NVIDIA infrastructure.
- Every command is logged with its result.
What IT teams inherit from building systems
Operational technology arrives in the building long before it arrives in the IT review.
Shadow OT
BMS controllers, meters and wireless sensors are bought by facilities, installed by contractors and found by IT later.
A vendor per system
Each building system has its own cloud, its own login and its own data model, none of which talk to each other.
Data locked in silos
Building data sits in proprietary historians, so the data team cannot join it with finance, occupancy or ERP data.
Data residency questions
Some sites, sectors and governments need the data and the AI to stay on premises.
What IT and data teams get from Aqueduct
One governed layer for building data, deployed the way your policies require.
Open protocols
BACnet, Modbus, MQTT, M-Bus, pulse and LoRaWAN, over LTE, Wi-Fi or Ethernet.
Learn moreCMMS and enterprise links
Tickets into your CMMS, data out to EAM, IWMS, historians and BI.
Learn moreQuestions IT and data leaders ask
“What certifications does Aqueduct hold?”
Milvian AI holds ISO/IEC 27001, ISO 9001:2015 and SOC 2 Type 2. Milvian AI is an AWS Advanced Tier Partner and an NVIDIA Inception member.
“Can it run entirely on our premises?”
Yes. Aqueduct can be deployed on-premises, including on-prem AI on NVIDIA infrastructure, where policy or data residency requires it.
“How does building data get into Aqueduct?”
LoRaWAN devices connect through gateways, IP devices over Ethernet, Wi-Fi or LTE, and BMS equipment through a field device network controller speaking BACnet or Modbus and publishing over MQTT.
“Can we get the data out to our own platforms?”
Yes. Aqueduct holds sites, devices and telemetry in one model and integrates with CMMS, EAM, IWMS, historians and other enterprise systems.
“Who can send commands to equipment?”
Control runs only where enabled, with owner approval. Lockouts block conflicting commands and command history records every command and its result.
Building data deserves the same governance as every other enterprise system.
Talk to our teamHow Aqueduct shows up in an IT team's month
An illustrative rollout across several buildings.
Week one
Security review: deployment model chosen, network paths for gateways and controllers agreed, roles and access set.
Week two
Existing BMS points and meters connect through gateways and field device controllers. Installation of new devices runs through partners.
Week three
The data team maps sites, floors and zones in aQ Core and connects the CMMS for ticket routing.
Week four
Device health shows connection rate per device, and the platform moves into normal operation under IT's change process.
Platform pages and industries IT teams read first
The technical depth sits here.
Where teams put it to work
Data and AI that stay on site
Public sector and healthcare sites deploy Aqueduct on-premises, with on-prem AI on NVIDIA infrastructure, so data never leaves their control.
Learn more
Power and cooling in one model
Data center teams bring feeder power, power quality and cooling data into aQ Core and integrate it with their existing tooling.
Learn more
Existing systems connected, not replaced
BMS points, meters and CMMS connect through open protocols, so IT approves one platform instead of a vendor per building system.
Learn moreWhat an IT review of building data actually covers
Facility systems arrive with network, identity, data and control questions that most vendors answer late. Here is how each one is handled before rollout, not after.
Where the traffic goes
LoRaWAN sensors talk to gateways, and gateways backhaul over Ethernet, Wi-Fi or LTE. BMS equipment is reached through a field device network controller that speaks BACnet or Modbus and publishes over MQTT.
- Backhaul paths agreed with IT before any device ships
- LTE backhaul keeps devices off the corporate LAN where policy prefers
- Outbound-only data paths available for read-only sites
- Inputs
- LoRaWAN, BACnet, Modbus, MQTT, M-Bus, pulse
- Transport
- Ethernet, Wi-Fi or LTE
- Aqueduct
- one agreed path per building
Where the data and the AI live
Aqueduct runs as a managed deployment on AWS or on your own infrastructure. Where data residency or sector rules require it, the AI runs on-premises on NVIDIA infrastructure.
- Cloud or on-premises, chosen per policy
- On-prem AI for regulated or sovereign deployments
- The same data model in both
- Options
- AWS cloud or on-premises
- AI
- on-prem on NVIDIA infrastructure where required
- Aqueduct
- deployed the way your policy reads
One model instead of one per vendor
aQ Core holds sites, floors, zones and devices as one hierarchy. Water, energy, air, controls and billing all read the same records, so a meter is defined once.
- Site, floor, zone and device hierarchy
- Device health per sensor: connection rate and battery or capacitor voltage
- Coverage gaps visible before anyone reports on them
- Structure
- sites > floors > zones > devices
- Quality
- connection rate per device
- Aqueduct
- one record per asset across modules
Who can change equipment, and how it is audited
Control is the part security teams ask about first. Aqueduct acts only where control is enabled, with owner approval, through two paths: LoRaWAN relays for simple loads and field device controllers for complex equipment.
- Read-only deployments for critical sites
- Lockouts block conflicting commands mid-action
- Command history records each command and whether it executed
- Paths
- LoRaWAN relays, BACnet or Modbus controllers
- Guardrail
- owner approval and lockouts
- Aqueduct
- full command history
How data reaches your own platforms
Building data is most useful joined with finance, occupancy and asset records. Aqueduct routes tickets into your CMMS and integrates with EAM, IWMS, historians and BI.
- Tickets with evidence into the CMMS you run
- Data out to historians, BI and the data warehouse
- aQ Assistant answers in context and flags data quality
- Out
- CMMS, EAM, IWMS, historians, BI
- AI
- aQ Assistant on platform data
- Aqueduct
- an integration surface, not a silo
The operating metrics an IT or data team watches
Building data has its own health signals. These are the ones Aqueduct exposes so IT can run it like any other governed system.
| Metric | Where it shows | Why IT cares |
|---|---|---|
| Connection rate per device | Device health tab in aQ Core | Shows which sensor is failing before reports go wrong |
| Battery or capacitor voltage | Device health tab | Plans replacement instead of site walks |
| Sub-metering coverage | Electrical twin footer and portfolio views | Tells data users how much of the total is attributable |
| Days with low coverage | aQ Assistant data quality notes | Stops week-over-week comparisons on incomplete data |
| Commands sent and executed | Command history per device | Audit trail for every change to equipment |
| Commands held by lockout | Command history | Evidence that guardrails are working |
| Meter-to-platform connectivity | Billing analytics per meter | Confidence level before tenant or supplier bills go out |
From security review to normal operation
An illustrative path for a multi-building estate. Timing depends on site count and how much existing metering already emits telemetry.
Architecture and policy
Deployment model chosen, network paths for gateways and controllers agreed, roles and access mapped to your identity policy, and read-only sites named.
Connect what exists
Existing meters and BMS points connect through gateways and field device controllers. Utility bills upload for day-one baselines.
Model and integrate
Sites, floors and zones are modeled in aQ Core. The CMMS is connected for ticket routing and data feeds are set up for BI.
Run under change control
Device health is reviewed like any other system. Control is enabled site by site, with owner approval, under your change process.
sites in 21 countries run on Aqueduct.
and 9 clinics in one of India’s top hospital chains.
US airport running water intelligence on Aqueduct.
of commercial real estate under Aqueduct.
FMCG manufacturing campus.
operations run on Aqueduct.
Related
Common questions
Milvian AI holds ISO/IEC 27001, ISO 9001:2015 and SOC 2 Type 2, and Aqueduct can run in the AWS cloud or on-premises. Control of equipment runs only where enabled with owner approval, and every command is recorded in command history.
Yes. Aqueduct can run on-premises, including on-prem AI on NVIDIA infrastructure, for organizations with data residency, sovereignty or air-gap requirements. Milvian AI is an NVIDIA Inception member.
Aqueduct supports LoRaWAN, Modbus, BACnet, MQTT, M-Bus and pulse outputs, with connectivity over LTE, Wi-Fi or Ethernet. It is device agnostic and connects anything that emits telemetry.
Yes. Aqueduct routes tickets into your CMMS and integrates with EAM, IWMS, historians and other enterprise systems. It does not replace your CMMS.
The cloud deployment runs on AWS, and Milvian AI is an AWS Advanced Tier Partner. Organizations that need data on their own infrastructure can deploy Aqueduct on-premises instead.
aQ Assistant answers questions from the data in your Aqueduct tenant, using the site or page you are viewing as context. Its answers include data quality warnings when coverage is low.
Not always. LoRaWAN devices use gateways that backhaul over Ethernet, Wi-Fi or LTE, and BMS equipment connects through a field device network controller. The network design is agreed with IT before installation.
Yes, in many cases. LoRaWAN gateways and controllers can backhaul over LTE, and read-only sites can use an outbound-only data path. The network design is agreed with IT before rollout.
Every device reports a connection rate, and battery devices report voltage. Coverage percentages show how much of a total is metered, and aQ Assistant notes when low-coverage days make a comparison unreliable.
Control runs only where it has been enabled, with owner approval, and sites can stay read-only indefinitely. Every command and its result are kept in command history for review.
No. Aqueduct holds the building data model and integrates with historians, BI and the warehouse, so building data can be joined with finance, occupancy and asset records.
See Aqueduct on your own facility.
Bring a site, a utility bill or a question. We will show you where the loss is.