A Diskos for the physical economy.

Make standardized, outcome-linked data capture a condition of support for projects advanced through Canada’s Major Projects Office.

The recommendation

Create a Canadian Project Outcomes Repository and require staged deposits from Major Projects Office (MPO) projects through approval conditions, contribution and loan agreements, procurement terms and—where necessary—sector regulation. Start with three pilots.

Design the obligation as capture once, report once: standardize data already produced by project controls, inspections, equipment, environmental monitoring and operations instead of creating a parallel reporting bureaucracy.

The opportunity

Construction sites, mines, power corridors and ports generate the physical-world evidence needed to improve safety, productivity, automation and robotics. Today, much of that evidence is fragmented across contractors, proprietary systems and short retention schedules.

Canada can preserve it under Canadian governance, use it to improve public delivery and license rights-cleared subsets commercially. Licensing is an upside and a cost offset—not a promise that data revenue will repay capital projects.

The scale makes a common standard valuable now: 27 MPO initiatives referred since September 2025, $200B in investment represented and a stated path to $500B in future private investment. Conventions adopted at the first projects become national infrastructure rather than expensive retrofits.

Diskos proves that mandatory submission can protect commercial value

Norway’s Diskos National Data Repository, operating since 1995, is the relevant institutional precedent. Petroleum licensees have a legal duty to report well data; most submissions flow through Diskos. The licensee remains responsible for completeness and quality, including data created by third-party contractors. Submissions carry structured metadata and are governed by confidentiality periods before release.

What Canada should copy

• A clear reporting duty, assigned to the project proponent even when contractors create the dataset.
• A data dictionary covering identifiers, provenance, units, timing, formats and quality.
• Tiered entitlements so regulators, rights holders, proponents and later users see only what they are entitled to see.
• Long-term stewardship, with version history and documented corrections.

What Canada should not copy blindly

• Automatic public release is unsuitable for every infrastructure, security or community dataset.
• One confidentiality clock cannot serve trade secrets, worker data, security-sensitive systems and Indigenous knowledge.
• Raw volume as success misses the point. The valuable unit is an event linked to a verified outcome.

The MPO can make the standard part of how projects proceed

The Building Canada Act framework is built around coordinated approvals and a document setting the conditions under which a listed project may proceed. That creates a practical policy window: data stewardship can be framed as a predictable delivery condition, not an extra assessment gate.

1. Set a default obligation. Every MPO-listed project submits a proportionate Data Capture and Stewardship Plan before major construction.
2. Attach it through the available instrument. Use the Act’s conditions document where legally available; otherwise use federal funding, loan, procurement or project agreements. Sector regulation remains the durable backstop.
3. Publish one national schema. Require common event IDs, timestamps, location, task, equipment, input conditions, intervention, outcome, quality score, rights status and access tier.

What the repository holds

The project record — design revisions and machine-readable as-builts; work packages, sequence and site conditions; equipment telemetry and maintenance events; material batches, tests and quality control; environmental measurements and mitigations; surveillance footage where a defined use justifies it.

The outcome layer — acceptance, defect, rework and failure labels; inspection findings tied to the underlying task; safety events and corrective actions; commissioning and operating performance; longitudinal condition and maintenance results.

The repository should preserve the chain from conditions to action to result. A useful record answers: what was attempted, under what conditions, with which tools and materials, what changed and how was the result verified?

Minimum-necessary collection: default to machine, process and asset data. Capture identifiable worker video, audio or biometrics only when a defined use justifies it, consent and labour rules permit it and a less intrusive method cannot deliver the same evidence.

Governance must precede collection

Rights and custody

The proponent retains its background intellectual property and protected commercial information. The repository receives a durable licence for sovereign preservation, regulatory use and approved research. Commercial reuse occurs only where every relevant right is cleared.

Indigenous data sovereignty

Data about First Nations people, communities, lands, waters, resources or knowledge cannot be treated as ordinary Crown data. Project-specific governance must be co-designed with affected rights holders. For First Nations data, the OCAP principles—Ownership, Control, Access and Possession—provide the starting framework, not a box-checking exercise.

Security and access

Keep the repository in Canadian-controlled infrastructure. Use tiered access: public, research, commercial, restricted and protected. Apply security review, export controls, audit logs, purpose limits and revocable access to sensitive infrastructure data.

Commercialization

License curated, rights-cleared datasets rather than selling unrestricted raw archives. Revenue should first fund curation and stewardship, then flow under transparent project agreements to relevant data contributors, rights holders, Indigenous partners and the public repository.

Three pilots can test the model across energy, mining and northern transport

18-month test: success measures

• 0–90 days: appoint an industry–Indigenous-government design table; select datasets; complete privacy, security and rights impact assessments.
• 3–9 months: publish schema v1; execute data agreements; ingest historical and live pilot data.
• 9–18 months: validate quality; test research and commercial access; publish results and a scale/no-scale recommendation.

Measures throughout: incremental capture cost per project; share of records linked to verified outcomes; time from event to validated deposit; rights-clearance and access turnaround; demonstrated safety, delivery or maintenance value; qualified demand and licensing revenue, if any.

Decision requested: direct the MPO, Treasury Board Secretariat, Innovation, Science and Economic Development Canada, Natural Resources Canada and Indigenous partners to produce a pilot mandate, legal pathway and common data standard within 90 days.