SiteIQ Position Paper VIC

In one paragraph

Victoria has publicly committed A$25 billion of investment attraction to data centres, and has announced a Sustainable Data Centre Action Plan that its own Legislative Council was told, on 13 May 2026, is “not downloadable, and multiple calls to ministerial staff and departmental contacts have resulted in either no greater clarity or a suggestion that it either has not been produced yet or is for internal government use only” (Hansard, 13 May 2026). One month later, the Attorney-General confirmed on the parallel water review that “no documents were identified that are responsive to the order” (Hansard, 16 June 2026). In the intervening ten weeks, the Lord Mayor of Melbourne joined a global mayors’ initiative on data centre burden (Reuters, 22 June 2026), the Greens and Family First separately called for moratoria, the South Morang community petition has passed 3,150 signatures (Parliament of Victoria petition, current as at 22 July 2026), and The Age editorial position on 22 July 2026 identifies Plumpton, West Footscray, Geelong and Port Melbourne as compounding community-alarm sites (The Age, 22 July 2026). Victoria does not have a data centre problem. It has a site-intelligence problem. This paper argues that the fastest, most defensible, and most fiscally responsible way to close that gap is to procure site intelligence through the existing DTP Geospatial Data and Analytics Panel. A Victoria-only site-intelligence tool is in build, with a demonstration release targeted for Q2 2026, a peer-reviewed methodology under discussion with competent experts across academia and industry, and a stated sequencing that anchors Victorian government relationships before any national or international expansion. The question is whether Victoria opens the category in weeks, or waits until the absence of a defensible instrument lands on the desk of the next Auditor-General.

1. The trilemma the Action Plan does not yet resolve

Three constraints, power, water and planning, each with its own department, its own regulator, and its own political frame, converge on every one of the roughly 40 existing and 21 projected large data centres in Victoria. No published Victorian instrument currently produces a site-level composite score across all three. DJSIR’s Action Plan is described on the department’s own website as “primarily an internal coordination framework” (DJSIR). Budget Paper 3 for 2026-27 funds “a framework for data centre investment and risk management” in future tense (Budget Paper 3). The DTP Geospatial Panel is the standing procurement rail (buyingfor.vic.gov.au), but there is nothing on it today that answers the composite site question. The result is the political timeline below.

The Commonwealth’s National Standards for AI Data Centres, announced on 15 July 2026 and targeted for legislation in early 2027 (Australian Government policy statement, 15 July 2026), sharpens rather than displaces the Victorian gap. National standards on energy, water, location and skills will apply to new proposals; states retain planning, water-corporation and site-approval functions, and will need a defensible site-level assessment layer to operate under those national standards from the day they commence. Victoria that has procured its own site-intelligence layer before the Standards commence will be the state that shapes how they are applied.

Political-pressure escalation, February to July 2026

Four categories of actor, being the executive, parliament, local government and community, have moved on data centres in the past six months. Every escalation has increased the political cost of not having a defensible, transparent site-selection instrument in the public domain.

2. Why the answer is procurement

Three properties of the problem determine what instrument fits it:

• The product is a data service, not physical infrastructure. Its value refreshes monthly. Its credibility rests on the market believing it is not captured by any single party, including government. Instruments designed for 20-to-30-year risk transfer on schools, hospitals and tollways were not built for a SaaS data product, and forcing that fit would produce contractual rigidity precisely where the methodology needs to keep moving.
• The score must be an input, not a decision. If the state produces the score directly, every rejected site’s developer has an administrative-law path against government at VCAT. If the score is procured from an arm’s-length data service, it remains evidence, and the decision remains with the department, the corporation or the tribunal. Procurement is procedurally cleaner.
• Timeline compression is the binding constraint. The Action Plan was announced in the 2025-26 budget cycle, described as an internal coordination framework twelve months later, and told to the Legislative Council in May 2026 as effectively non-existent as a public document. Building an equivalent capability inside government from scratch, including data model, ETL, methodology, review, audit and publication, has historically taken 18 to 36 months and 5 to 10 times commercial cost. Every additional month is another South Morang petition, another Plumpton town-hall, another Family First press conference.

The concern that composite scores over Crown data should not become a private rent is real and must be addressed on substance. It is addressed in §4 through five specific governance obligations that are stronger under procurement than under any alternative structure. Those obligations are the difference between rent-extraction and public-interest procurement, and they are what make this a procurable category, not a bespoke instrument.

The three routes compared

3. What Victoria would actually procure

A site-intelligence data service with the following minimum published characteristics. SiteIQ is proposed as one candidate under the DTP Geospatial Data and Analytics Panel. This paper is not the procurement, it is the case for opening the category.

Coverage. Every large-load-capable land parcel in Victoria, updated monthly, with a site-level composite score on a 1 to 10 scale across seven pillars: Power, Water, Fibre and Connectivity, Planning, Physical Climate Risk, Fossil-Gas Alignment, and Social-Licence Context. Weights transparent, versioned and adjustable per department use case. Underlying primary sources include AEMO published grid data, DEECA Victorian Water Accounts, Vicmap Planning, published carrier point-of-presence inventories, BOM and open climate-trajectory datasets, and standard demographic and community datasets. Every value is stamped with its data source, retrieval date and methodology version.

Derived data assets. Beyond the composite score itself, the service generates a set of derived spatial data assets not currently in the public domain: cumulative water-catchment demand overlays, non-potable feasibility layers, cluster-level grid-headroom views, and time-series site-level trend files. Under the community-facing layer commitment below, non-commercially-sensitive derivatives of these assets are contributable back to data.vic.gov.au and Vicmap. This converts the procurement from a scoring layer into a data-supply pathway back into Victorian public data infrastructure.

Methodology. Published, versioned, peer-reviewed. Independent methodology review across each pillar is under discussion with competent experts across academia and industry. The water pillar methodology is in v0.9 draft. Proposed publication venue is a recognised peer-reviewed venue appropriate to each pillar. The methodology explicitly complements, and does not substitute for, the DEECA Industry Water Connection Guide, the DEECA/VicWater expert review, and existing statutory heritage and planning regimes.

Data outputs to government. Score, sub-scores, confidence flags, cumulative-impact catchment and cluster views, provenance and methodology version on every value. Machine-readable feeds under the Victorian Government API standards.

Community-facing layer. An anonymised, site-non-identifying public register of what the score would be if a hypothetical data centre were proposed on a given zoned parcel. This is the transparency function that the Ettershank motions, the De Vietri matter, the South Morang petition and the Family First position all implicitly demand. It is the piece the market cannot deliver on its own, and that procurement under a public panel can compel.

Explicit exclusions. The tool does not replace water corporation servicing assessments, VCAT decisions, or planning permit determinations. It is an evidence base, not a decision-maker. Nor does it substitute for statutory functions of DJSIR, DEECA, water corporations, DTP, or the Victorian Auditor-General. Where the tool integrates with those functions (for example through data-supply agreements with water corporations), those integrations are formally contracted and separately audited.

4. Governance obligations that make procurement the more accountable path

The public-interest concern with any private product over Crown data is legitimate. It is addressed procedurally, not structurally:

• Peer-reviewed methodology publication, revised annually, with independent external review across each pillar drawn from competent experts across academia and industry, including explicit reference to which regulatory frameworks the methodology complements and which it does not substitute for.
• Open data feedback obligation. The provider must contribute derived, non-commercially-sensitive layers back to data.vic.gov.au and Vicmap where feasible.
• Right of methodology audit. The Auditor-General and any nominated departmental steward may audit the methodology and data pipeline. The Victorian Auditor-General’s 2026-27 Annual Plan already flags data centre oversight as an emerging area of interest (VAGO Annual Plan 2026-27).
• Panel-exit rights. A standard DTP Panel contract is terminable and re-competable. There is no lock-in on either side.
• Living-methodology commitment. Annual methodology refresh aligned to the Victorian Water Accounts water year, with review-panel sign-off before publication.

These five obligations make procurement, on any reasonable test, the more publicly accountable path, not the less.

5. Who this paper is for, and what SiteIQ answers for each

This paper is a first introduction to SiteIQ. It is not a first introduction to the underlying policy question, which is well ventilated on the Hansard record and in the community campaigns cited above. Each recipient has a specific problem that a procured site-intelligence layer directly answers.

6. What we are not asking

• We are not asking for grant funding, equity, an exclusive dealing arrangement, or a bespoke structural instrument.
• We are not asking government to endorse SiteIQ ahead of a competitive Panel procurement.
• We are not asking for expedited planning, water or grid treatment for any site the tool scores well.

We are asking government to open a category, being site intelligence for large-load infrastructure, under an existing panel, and to let the market, the methodology reviewers, and the audit right in §4 decide what fills it. We note separately that discussions with individual water corporations on data-supply MoUs are commercial arrangements outside the scope of this procurement submission and are not the subject of this paper.

7. Closing note on the metaphor

An albatross around the neck is not an insult to the bird. It is a burden the wearer chose to keep. Victoria’s Sustainable Data Centre Action Plan is not the burden. The absence of a site-intelligence layer under it is the problem to be solved -the layer that lets a water corporation, a planner, a mayor and a hyperscaler see the same number and argue over its inputs rather than its politics. The bird is small. The wearer can safely remove it. This paper is the instruction manual.