CANUS
A PortaDC vault arriving past rows of solar panels and battery cabinets

CANUS/Technology / Solar-Sponge™

More compute when solar is abundant.Less demand when the site needs power.

Proprietary thermal-mass buffering and workload-shaping architecture that absorbs high-output renewable generation windows and rides through transient cloud or tariff events without evaporative water or grid stress.

All four systems

01In operation

01 / 04

01In operation01 / 04

Solar-Sponge™Solar-Synced Scheduling & Thermal Buffer

Modelled 24-hour operating day (900 kW cluster · 2 × 450 kW pods · 1 MW boundary)

Synchronised: cluster demand above · cushion state below

1 MW SITE BOUNDARYCLUSTER DEMAND · 900 kW IT FULL COMPUTESURGE BANDCUSHION STATE OF CHARGE (%)00:0004:0008:0012:0016:0020:0024:00

12:00 · Full compute on daytime solar · cushion charging

Cluster demand

900kW

Token output

100%

Cushion charge

84%

  • 00:00–08:00 & 20:00–24:00

    Overnight grid & battery dispatch: hybrid tariff passed through at cost under the site schedule

  • 08:00–17:00

    Daytime behind-the-meter solar window: full compute, cushion charging on renewable output

  • 17:00–20:00 · when called

    Host call: the cushion is released at the boundary; Surge tenants shape their own contracted band, Pulse untouched

450 kW → 1.2 MW
Per Pod, Two Hulls
22.5 MW+
Per Campus, Run as One Fleet
120 days
Target from Order to Powered Racks
-12–21%
Under Sydney Metro, All-in
0 L
Evaporative Cooling Water
A PortaDC vault arriving past rows of solar panels and battery cabinets

Solar-Sponge™Solar-Synced Scheduling & Thermal Buffer

02What it does02 / 04

What it does.What it means.

  1. 01

    The power ledger

    Power at cost, one line

    What it does

    Direct behind-the-meter renewable and hybrid tariffs passed through under transparent site schedules (shared under NDA), delivering 12 to 21% lower all-in compute cost than Sydney metro colocation. Power is passed through at cost, not marked up.

    What it means

    Rent and power sit on separate lines of one invoice, so a tenant can trace each, and the solar behind it is recorded hour by hour.

  2. 02

    The cushion, planned

    Hour by hour

    What it does

    During midday solar, Solar-Sponge™ plans the flexible band around the solar day. It plans the cushion hour by hour; FrontierGate™ executes it at the boundary.

    What it means

    The host's evening headroom comes from the cushion, never from the tenant's firm floor. The GPUs never notice.

  3. 03

    Surge only: the tenant-authorised flexible band

    25 to 40%, contracted, tenant-run

    What it does

    Tenant-authorised and Surge only: when the host calls, Solar-Sponge™ publishes a power-budget signal inside the contracted 25 to 40% band and the tenant's own orchestration answers it, shaping its batch, fine-tuning and prefill queues however it chooses. Pulse tenants have a band of zero and are never trimmed.

    What it means

    The retention a Surge tenant holds through a called event is its own target, set by its orchestration, not a CANUS promise. The host meets its flexibility obligation; the tenant's software is never touched.

03In detail03 / 04

Indetail.

The power ledger

Passed through at cost, separate from the rent · site tariffs shared under NDA

  • Daytime solar

    Behind the meter · 08:00–17:00 · site schedule under NDA

    Direct

  • Blended 24 h

    Solar + overnight grid & battery dispatch · passed through at cost

    Hybrid

  • Metro hall

    Retail grid mix plus network charges, not passed through at cost

    Tariff

All-in, Pulse

21% under

Against Sydney metro colocation, all-in

Carbon

Scope 2, matched

Interval-matched proof of the solar behind the tokens

All-in cost vs the hall

Target rent + power at cost · US$ per IT-kW-month · you bring the GPUs · 900 kW cluster

PortaDC Surge™

~US$182 / IT-kW / mo

Sydney metro, all-in

~US$270 / IT-kW / mo

Under Sydney metro

33%

Against the hall, per year · 900 kW

US$0.95M

Indicative figures, confirmed in each lease. Site tariffs are shared under NDA. You own the GPUs on both sides; the gap is the rent and the power.

Deploy

Tell us the workload.We'll bring the pod.

A short conversation about what you run, where your customers are and when you need it.

The pod

Up next

LARUS AI™→