Start with a pilot. One facility, 30 to 90 days, with clear success metrics. Then scale per megawatt.

    Get started with a pilot

    Build your estimate

    How AI data centers buy electrical intelligence

    A subscription for most AI-DC fleets, a CapEx license for finance teams that need to own it, and an audit-grade tier for regulated workloads.

    Trusted by leading AI data center operators

    NVIDIA logo
    AWS logo
    Google logo
    Arm logo
    Qualcomm logo
    T-Mobile logo
    Digital Realty logo
    STACK Infrastructure logo
    Westinghouse logo
    NVIDIA logo
    AWS logo
    Google logo
    Arm logo
    Qualcomm logo
    T-Mobile logo
    Digital Realty logo
    STACK Infrastructure logo
    Westinghouse logo

    Deployment estimator

    Build your estimate in four steps

    No email required. Configure your deployment and watch the estimate update live.

    Choose how you buy

    Size your fleet

    Pick your usage profile

    Anomaly detection, predictive maintenance, and workflow automation ramping over the term.

    Commitment & support

    24/7 support included. Committed-use discounts lock your rate and lower the effective cost.

    Your estimate
    $238K / yr avg
    SKUCloud (OpEx)
    Base subscription$135K/yr
    ProfileMid
    CommitmentNone
    Indicative 5-yr TCO$1.19M
    What you get
    • 8 kHz waveform + ML fault detection
    • Hardware bundled & financed, 24/7 support
    • Telemetry + 1-yr retention + included credit pool
    • Verdigris-managed platform & uptime

    Order-of-magnitude estimate. Not a quote. Final pricing scoped on a deployment call with our engineers.

    Cloud fits teams funding from operating budgets who want Verdigris to run the platform.

    Get started with a pilot

    Procurement FAQ

    What's the pricing basis?
    We charge by deployed capacity in megawatts. Cloud starts at $25,000 per MW per year. For pilots or sites under 1 MW, we charge per gateway, then switch to per-MW pricing once you cross about 50 gateways or 2 MW deployed. We share specific numbers on a scoping call after we see your fleet. T-Mobile measured 6:1 ROI on 800 rectifiers.
    Can we mix SKUs across facilities?
    Yes. A common pattern is Cloud across the operating fleet, Forge for a new CapEx build, and Sentinel for one regulated workload. A single Verdigris contract covers all three.
    How does Verdigris integrate with our existing BMS, DCIM, and ticketing systems?
    Verdigris becomes your single source of truth for electrical state. It samples at 8 kHz, while a BMS samples at 1 Hz, so it sees things they miss. Your existing Schneider, Honeywell, Eaton, or Vertiv gear keeps measuring and feeds data in over BACnet, Modbus, REST, or MQTT. ServiceNow can auto-create work orders from Verdigris alerts. Most brownfield sites install without custom integration work, because we ingest whatever protocol you already use.

    Full API and webhook documentation is at docs.verdigris.co. No NDA required.
    How does this map to our CapEx vs OpEx categories?
    Cloud is an OpEx subscription, the same line item as your DCIM and observability tools. Forge is CapEx: a perpetual license plus hardware your team owns, with a separate OpEx line for maintenance. Sentinel is an OpEx subscription too. Pick Forge if your finance org needs the platform on the balance sheet. Pick Cloud if you fund from operating budgets.
    What's the minimum commitment?
    Pilots run with no commit: one facility, 5 gateways, 30 to 90 days, with clear success metrics. Cloud is a per-MW subscription with annual or multi-year terms once you scale past about 50 gateways or 2 MW deployed. Forge is a perpetual license plus annual maintenance, scoped to your CapEx cycle. Sentinel is a subscription with custom contract terms for regulated environments.
    Can we add Sentinel later if a regulated workload comes online?
    Yes. Sentinel covers a specific workload, not a tier above Cloud. Most customers add it for one regulated AI workload while Cloud keeps running the rest of the fleet. Cloud and Sentinel run together under a single contract. They are not upgrade or downgrade tiers.
    What's the deployment timeline?
    Timing depends on your fleet size, the systems we need to connect to, and SKU choice. Cloud is fastest, since Verdigris manages the platform end to end. Sentinel runs about four weeks with a dedicated solution architect, to meet audit-grade compliance. Forge takes longest, because your team owns install, upgrades, and on-call. We commit to a date range once we have a clear picture of your site.
    Who owns the hardware in Forge?
    You do. Forge is a perpetual license, so your team owns the hardware on the balance sheet and the software license. Verdigris provides maintenance and platform updates under an ongoing service agreement.

    Want a hand in what Signals detects? Signals Founding Access is a sponsored co-development cohort: Verdigris funds the hardware and the engineering, you bring the floor and help define the signals before launch.

    See Founding Access

    What happens after you reach out

    Our engineers walk your single-line diagrams with your team, scope the deployment, and agree on what a successful pilot looks like before we commit to a date. You see results in your facility before any decision to scale.

    Proven at fleet scale

    T-Mobile measured 6:1 ROI across 800 rectifiers. Verdigris caught degradation in 4% of the fleet 21 days before any standard alarm fired.

    Audit-ready

    SOC 2 Type II certified

    SOC 2 Type II certified. Regulated environments welcome. Sentinel adds a dedicated solution architect and audit-grade compliance.