Platform architecture
Domain fabrics on shared foundations, all under one governed plane.
Datacenter · DCIM · GPU cloud
SynapseDC: plan, build and operate datacenters on one platform
One estate model from the utility feed to the GPU. Plan the next hall, build it under change control, and run it knowing who every fault touches and what it costs.
What’s inside.
- Colocation
- GPU cloud
- Facility power & cooling
- GPU fleet
- Revenue at risk
- Capacity & power economics
- Datacenter planning
- DCIM
Plan
- Templates for halls, sites and plant
- Sizing, CapEx, NPV and build schedule
- Approvals by plan type and spend
Build
- Racks, power and cabling as records
- Every edit tagged to a change request
- Work orders and L1–L5 commissioning
Operate
- Alarms scoped by site and dependency
- Power path, UPS and cooling redundancy
- GPU jobs, blast radius and time lost
Money at risk
- Revenue lost and SLA credits per issue
- Sellable capacity, not just free racks
- Idle contracted kW, cost per GPU-hour
Data & deployment
- Read-only collector, outbound mTLS
- BMS, power, cooling, BMC and GPU data
- Cloud, on-premises and air-gapped
Built for the datacenter, end to end.
Estate health
Command Center
Every datacenter on one screen, ordered by what matters, with the live map and what to do first.
Revenue impact
Revenue Impact
Which issue is costing the most right now: money lost, SLA credits and the next credit tier.
AI infrastructure
GPU Fleet
One failed GPU stops a whole job. See which GPU, whose job and what the lost GPU time costs.
Plan & build
Planner & DCIM
From a requirement to an approved plan pack, then built and recorded under change control.
What you see on day one.
Who
Customers at risk
Tenants and services on the power path, cable or GPU job, premium first.
Colocation · NeoCloud
How much
Money on every issue
Revenue lost so far, SLA credits by tier and the cost of waiting another hour.
Revenue Impact
How long
Ride-through, in minutes
UPS minutes left, diesel fuel hours and the cooling spare left on each floor.
Facilities
What next
Capacity you can sell
Sellable racks and kW, the next hall’s cost and go-live date, kit past end of sale.
Capacity · Planner
Early deployments are being baselined with colocation and GPU cloud operators. Results depend on your estate; we validate against your own baseline.
Where SynapseDC leads.
Datacenters & AI infrastructure: colocation providers and GPU clouds first, then cloud, enterprise and telco datacenters.
In the operation.
How it fits.
SynapseDC runs on the VisionWaves platform, under the same governed plane as every other fabric. It uses Opstral AI and the platform’s modules through their APIs, and reads the systems already in your datacenter rather than replacing them.
Governance
Every disruptive step waits for a named approver; plant control stays with your BMS.
Integrations
Reads BMS, power, cooling, BMC and GPU data; works with your ITSM, CRM and IPAM.
Does SynapseDC replace our DCIM or BMS?
No. It reads from the BMS and the plant with a read-only role and never controls equipment. Its DCIM can be your record of the physical estate, or sit over the one you have.
Is SynapseDC only for operations?
No. It covers planning (capacity, new halls, refreshes), the build (DCIM, projects, commissioning) and operations, on one estate model.
Which business models does SynapseDC cover?
Colocation and GPU cloud (NeoCloud) in the first release; cloud, enterprise and telco datacenters follow on the same estate model.
Can SynapseDC run air-gapped?
Yes. It deploys on-premises, in your cloud or air-gapped. The on-site collector only ever connects outbound.
Codify judgement once. Compound it everywhere.
Pick one site and two situations your team keeps dealing with. We’ll connect it, set the value ambition with you and baseline it together.