EdgeQub

Practice 01 / Edge AI Campuses

AI capacity has to live somewhere.

The question is whether the infrastructure needed to carry it exists there too.

An Edge AI Campus establishes AI capacity around the conditions of the place where that capacity is needed.

One Campus drawn in section on real ground. Energy stands to one side, the engineered envelope sits on the ground profile, and the conditions of the site are marked around it.
The same section recomposed for a narrow viewport. Energy above, the engineered envelope on its ground, and site conditions marked below.

The energy that is there · the engineered envelope · the conditions of the site

Representative section

Representative depiction. The drawing carries no site, dimension, topology or distance, and no deployment is depicted.

The site decides what the infrastructure has to survive.

Category position

Australia's first Edge AI Campus

CAMPUS arrival positions

The capacity is available. The infrastructure that carries it is not.

The requirement may be additional capacity, capacity sooner, capacity somewhere conventional development would not go, or capacity where none exists today.

Extend. Bridge. Reach. Establish.

These are different positions. Not different products.

Extend

Capacity exists, and it is not where the next requirement is.

Capacity already exists in one position and the next requirement sits somewhere else.

Bridge

Capacity is needed now, and the programme that delivers it does not land for years.

Capacity established now, standing beside a longer programme that lands later.

Reach

The work is somewhere conventional development will not go.

Ground that conventional development does not reach, with capacity established beyond it.

Establish

AI capacity is being created where none exists today.

Capacity created where none exists, with its own energy standing beside it.

Representative positions. Not exhaustive, not exclusive, not products, and not a segmentation of any market or industry.

CAMPUS operating conditions

What the architecture is engineered against.

Every one of these is present at every site. What changes is which of them decides the design.

The architecture is engineered to the supply, the ambient and the access that are actually present, rather than to an assumed connection and a catalogue enclosure.

Power that varies by site and by hour

Ambient heat, and how much can leave

Compute density in a small envelope

Links that degrade and return

Environmental exposure

Integrity the operation depends on

A deployment horizon shorter than an infrastructure programme

The seven operating conditions, all present at every site. Representative, with no severity or priority asserted and no site described.

CAMPUS engineering substrate

The parts are available. The resolution is the product.

An Edge AI Campus resolves them as one Engineering Substrate.

Buying the four separately leaves somebody holding the interfaces between them, on every site, for the life of the asset. EdgeQub holds them instead, because they were engineered together.

A change in one can create consequences for the others. EdgeQub engineers the relationships between them as part of the Campus itself.

Power · Thermal · Compute · Communications

Power

supply that varies, held steady

Thermal

heat that has to leave with it

Compute

density that produces the work

Communications

links that degrade and return

Resolved against each other, not specified separately

One engineered body. The division is analytical, not physical. Representative depiction, and not a set of separately procurable products.

Look closer at one relationship. The body does not come apart.

Supply that varies, held steady.

AI capacity is an energy decision before it is a compute decision. Energy resource, storage and management can determine what capacity can be established and how it can operate.

Views of one engineered body. Selecting a view does not separate it, and none of the four is separately procurable.

CAMPUS thermal engineering

Compute density only matters if the heat can leave with it.

Thermal engineering is part of the Campus architecture from the beginning.

IceQub resolves the thermal path within the Campus rather than treating cooling as a later selection.

A compute envelope with the heat path drawn as part of the same engineered body, leaving the envelope into the ambient conditions of the site rather than being added afterwards.

Left · the compute envelope Right · the heat path, engineered with it

Below · the ambient the site actually has

Sectional and indicative. Not a physical layout, not a cooling topology, and no temperature, efficiency or thermal-performance value is asserted.

Thermal is not selected after the compute envelope is fixed.

It is engineered with it, against the ambient the site actually has.

CAMPUS engineered operating envelope

The architecture stays the same. The envelope is engineered.

EdgeQub standardises where the architecture needs to remain coherent and engineers where site conditions demand it.

The Engineered Operating Envelope establishes the conditions the Campus has been engineered to operate within.

Two bands. The upper band is the architecture, drawn as one constant. The lower band is the engineered envelope, where two different sites land in different positions.

Above · the architecture, constant

Below · the envelope, where a site lands. Filled mark, this site. Open mark, another site.

Marks indicate position, not measurement. No scale, value, range limit or site is asserted or described.

CAMPUS campus forms

Three forms. One Campus architecture.

Complete forms. Not stages.

Each is engineered to its own capacity. A Wing is not a step toward a Block.

The requirement determines the form.

Wing

500kW · engineered capacity

Campus Wing drawn complete on the shared module lattice, set against an environment that already exists.

Is an Edge AI Campus

Adds meaningful AI Campus capacity within an existing environment.

Block

1MW · engineered capacity

Campus Block drawn complete on the shared module lattice as a closed repeatable arrangement.

Is an Edge AI Campus

A repeatable building block of Edge AI Campus infrastructure, and Campus infrastructure in its own right.

Precinct

~100MW · engineered capacity

Campus Precinct drawn complete on the shared module lattice as several engineered groups held inside one architecture.

Is an Edge AI Campus

Edge AI Campus capacity at substantial infrastructure scale.

Common to all three

Each one is an Edge AI Campus · Engineering Substrate · Engineered Operating Envelope

Capacity anchors are stated because each form is engineered to a known capacity. They are not a range to shop across. Selected against the requirement. Not a sequence, not a prerequisite, not an upgrade path, and not a small, medium and large packaging. The three scenes are drawn independently and assert no relative physical scale between the forms. The reading order is a reading order only.

CAMPUS campus architecture

Architecture defines the Campus. Not the fence.

A Campus can occupy multiple physical positions and remain one coherent engineered architecture.

Distribution changes the relationships that need engineering. It does not remove them.

A wide elevation. Engineered capacity stands in three positions across the ground with one architecture bracket spanning all of it. No perimeter is drawn.
The same relationship recomposed vertically. Three engineered positions down the drawing with one architecture bracket holding all of them.

Engineered relationships, not a perimeter

Left · the energy that is there Right · the operation it exists to serve

Positions, energy and operation shown are representative. No number, distance, topology or latency is asserted.

CAMPUS energy precedence

Energy first. Everything else follows it.

AI capacity is an energy decision before it is a compute decision.

Energy resource, storage and management can determine what capacity can be established and how it can operate.

Energy resource, storage and management drawn as the three inputs that decide what AI capacity can be established, with capacity drawn to their right.

Resource · Storage · Management

Right · the capacity that can be established, and how it can operate

Representative depiction. No quantity, efficiency, emissions or environmental claim is asserted.

CAMPUS lifecycle integrity

Critical by design, or critical by accident.

Integrity has to survive the lifecycle.

When the conditions change, the engineering has to remain understandable.

One continuous obligation across five stages of the life of a deployment, with change returning to design rather than restarting the obligation.

Design · Build · Deploy · Operate · Change

One continuous obligation, not a delivery methodology and not a maturity model. Change returns to design. The obligation does not restart.

CAMPUS qub-space

A Qub is rack space inside an Edge AI Campus.

Qub-space is the rack space an Edge AI Campus makes available. Qubs are where compute physically stands once the Campus around them has been engineered.

The Campus is the engineered capacity. The Qubs are inside it.

An engineered Campus boundary with the substrate resolved around it. Inside it, a row of rack positions. The rack positions are Qub-space. The engineering around them is the Campus.
The same relationship redrawn for a narrow viewport. The engineered Campus boundary outside, rack positions inside.

Outer boundary · the engineered Campus

Inner positions · Qub-space

Representative depiction. Rack space is what a Qub is, not what an Edge AI Campus is. Power, thermal, communications and integrity are engineered around the Qubs and are not separable from them.

Rack space you can occupy. Because of everything engineered around it.

CAMPUS category questions

Category questions

You now understand the proposition. Here is what a new category asks.

Readers arrive carrying assumptions from conventional infrastructure. These resolve the ones that matter.

01What is an Edge AI Campus?

Engineered infrastructure that establishes AI capacity around the conditions of the place where that capacity is needed.

Power, thermal, compute and communications are resolved against each other as one Engineering Substrate rather than procured separately. That resolution is what EdgeQub engineers.

02How is it different from a conventional data centre?

Conventional development answers demand in large, concentrated, long-horizon increments, at sites chosen for that model.

A Campus is engineered to the supply, the ambient and the access actually present where the capacity is needed. It is also defined by its architecture rather than by a boundary fence, so capacity can stand in more than one position and still be one Campus.

03What is a Qub, and what is Qub-space?

A Qub is rack space inside an Edge AI Campus. Qub-space is the rack space an Edge AI Campus makes available.

Qubs are where compute physically stands once the Campus around them has been engineered.

04Is Qub-space just colocation?

Rack space is what a Qub is. It is not what an Edge AI Campus is.

Power, thermal, communications and integrity are engineered around the Qubs and are not separable from them. The Campus is the engineered capacity. The Qubs are inside it.

05Are Wing, Block and Precinct upgrade stages?

No. Each one is already an Edge AI Campus. Wing is engineered to 500 kW, Block to 1 MW and Precinct to approximately 100 MW.

They are not a sequence, not a prerequisite, not an upgrade path and not a small, medium and large range. A Wing is not a step toward a Block.

06Does the customer choose the form?

The requirement determines the form. Capacity is selected against the requirement, and the architecture, the substrate and the assurance basis do not change with it.

Configuration follows the engineering. Never precedes it.

07Why does EdgeQub begin with energy and thermal conditions?

AI capacity is an energy decision before it is a compute decision, and compute density only matters if the heat can leave with it.

Selecting cooling after the compute envelope is fixed is how sites end up limited by the one thing nobody engineered for them.

08Can a Campus exist without an Edge AI Factory?

Yes. Campus and Factory can stand independently.

Where the requirement calls for both they can be engineered together, but neither is a prerequisite of the other.

09Why is sovereignty more than physical location?

Control has to survive a dependency changing. Sovereignty becomes real when a dependency is tested, not when a location is claimed.

Parts of a capability can sit in different places without control becoming ambiguous.

CAMPUS commercial entry

Request Qub-space.

Tell us where you need AI capacity.

Request available Edge AI Campus capacity, bring us a site requirement to qualify, or register demand for future Campus capacity.

You do not need to know the physical configuration first.

Request Qub-spaceDiscuss a Campus requirement