Year 1 Prototype Priorities

Twenty-one prototypes to demonstrate NORAI capabilities and establish Canada as a global AI leader: including 6 quick-wins and 7 world-leading innovations

12-Month Development Cycle $144.5M Prototype Budget

Quick Wins (6 Prototypes)

High-impact projects deliverable within 3-6 months

Quick Win

Compute Queue Predictor

ML model predicting wait times across Alliance clusters, suggesting optimal submission times and resources, with auto-retry for failed jobs.

Q1 2026 (3 months) $1M
Quick Win

BorealisChat

RAG-based bilingual (EN/FR) chatbot trained on Canadian scientific literature, Tri-Council funding guidelines, and government reports. Hosted on 100% Canadian infrastructure.

Q1-Q2 2026 (6 months) $4M
Quick Win

Carbon-Aware Scheduler

Route batch jobs to lowest-carbon regions (Quebec hydro vs Alberta gas). Time-shift non-urgent jobs to low-carbon periods. Carbon footprint reporting.

Q1-Q2 2026 (4 months) $1.5M
Quick Win

Grant Writing Assistant

AI assistant trained on successful Tri-Council grants (anonymized), understanding NSERC/CIHR/SSHRC formats, reviewer criteria, and common pitfalls.

Q1-Q2 2026 (4 months) $2.5M
Quick Win

Research Impact Tracker

Automated tracking of dataset citations, code reuse, downstream discoveries, policy citations, and media mentions.

Q1 2026 (3 months) $2M
Quick Win

Collaboration Matcher

Knowledge graph of researchers, publications, datasets, grants with AI-powered similarity matching and collaboration recommendations.

Q1 2026 (3 months) $1.5M

Core Infrastructure (8 Prototypes)

Computome: The Compute Connectome

Unified gateway and intelligent dispatcher for all Digital Research Alliance resources. Automatic routing based on queue times, data locality, and carbon footprint.

Problem: Students wait weeks for campus clusters while national resources sit at 60% utilization. Startups pay $3/GPU-hour on AWS but cannot access Alliance resources.

Outcome: National cluster utilization increases from 60% to 80% ($50M value/year). Startups access HPC at $1/GPU-hour. Students get same priority as professors.

Q1-Q2 2026 $8M

Federated Data Catalog

Unified search API across all Canadian scientific repositories with FAIR metadata and OCAP compliance. Single interface to discover datasets across 13 provincial and federal systems.

Problem: Researchers spend 40-60% of their time searching for datasets across fragmented provincial and federal systems.

Outcome: Data discovery time reduced from weeks to minutes. 100+ datasets searchable in first release.

Q1-Q3 2026 $6M

Consent-as-Code (OCAP Engine)

Machine-readable consent policies with blockchain-based immutable audit logs. Community-controlled dashboards for Indigenous data governance.

Problem: No existing platform automates Indigenous data governance (OCAP) at scale. Manual compliance is slow and error-prone.

Outcome: First platform worldwide to automate OCAP at scale. Competitive advantage over US Genesis initiative.

Q2-Q4 2026 $7M

Provincial Data Sharing Protocol

Automated governance framework with legal compliance built-in for PIPEDA, provincial health acts, and OCAP requirements.

Problem: Cross-provincial research requires 6+ months of legal paperwork and manual compliance checks.

Outcome: Cross-provincial projects go from 6 months bureaucracy to 2-week digital workflow.

Q2-Q4 2026 $8M

Critical Minerals AI

AI-powered geological analysis combining satellite imagery, geophysical surveys, and historical data to identify economically viable rare earth and lithium deposits.

Problem: Canada needs to identify and develop critical mineral deposits to support clean energy transition and reduce dependence on foreign supply chains.

Outcome: 10 new deposit candidates identified. $15B+ potential export revenue. Support for clean energy supply chain.

Q2-Q4 2026 $9M

PermafrostGPT

High-resolution predictive modeling for northern infrastructure resilience. Machine learning models trained on 40+ years of ground temperature data.

Problem: Northern infrastructure faces billions in damage from permafrost degradation. Current monitoring is sparse and reactive.

Outcome: 1-km resolution forecasts. $8-12B projected savings in infrastructure protection across NWT and Yukon.

Q3 2026 - Q1 2027 $7M

Privacy-Preserving Health Analytics

Federated learning framework enabling cross-provincial cancer research without centralizing patient data. Secure multi-party computation across provincial data enclaves.

Problem: Provincial health data cannot be centralized due to privacy laws, blocking pan-Canadian research collaboration.

Outcome: Analysis of 10M+ patient records with zero data centralization. 3 novel therapeutic targets identified.

Q3 2026 - Q2 2027 $5M

NORAI Data Gateway

Unified API wrapper federating 30+ Canadian open science platforms (CONP, Ocean Networks Canada, CADC, Polar Data Catalogue, CanDIG, FRDR, and more) with native PyTorch/TensorFlow integration for AI/ML pipelines.

Problem: 30+ Canadian open science platforms exist but are siloed. Researchers waste 40-60% of time on data wrangling. No unified way to feed Canadian data into AI/ML pipelines.

Outcome: Single API to discover, access, and stream data from all platforms. Direct DataLoader integration for ML training. Automatic format conversion (DICOM, NetCDF, VCF to Parquet/Zarr).

Q1-Q4 2026 $8M

World-Leading Innovations (7 Prototypes)

2 World Firsts, 5 Global Competitive Advantages

World First

Arctic Digital Twin (Borealis-Earth)

First km-resolution digital twin of Canadian Arctic and boreal ecosystems, integrating permafrost, sea ice, wildlife corridors, and infrastructure. No equivalent exists globally.

Problem: Canada has 40% of the Arctic but no dedicated high-resolution digital twin. NVIDIA Earth-2 and EU DestinE focus on global/European climate.

Outcome: Enables predictive infrastructure planning, supports Arctic sovereignty, powers climate adaptation policy. Positions Canada as Arctic AI leader.

Q1-Q4 2026 (12 months) $12M
Genesis Parity

Autonomous Materials Discovery Lab

Self-driving laboratory for critical minerals and clean energy materials. Closed-loop AI + robotics: AI designs experiments, robots execute, ML analyzes, loop repeats 24/7.

Problem: US DOE Genesis Mission is building autonomous labs at Argonne/Berkeley. Canada has no equivalent. Manual materials discovery takes 10-20 years per new material.

Outcome: 10x faster materials discovery. Reduces critical minerals import dependence. Positions Canada in global battery/clean energy race.

Q1-Q4 2026 (12 months) $15M
World First

Indigenous AI Governance Platform

Complete technology stack for Indigenous data sovereignty: community-controlled consent portals, automated CARE/OCAP compliance, benefit-sharing smart contracts, data repatriation tools.

Problem: No platform anywhere automates Indigenous data governance at scale. CARE Principles and OCAP exist as frameworks but require manual implementation.

Outcome: World's first: exportable to Indigenous communities globally. Positions Canada as ethical AI leader. Addresses reconciliation commitments.

Q1-Q3 2026 (8 months) $5M
Sovereign AI

Borealis-Science Foundation Model

Sovereign 70B+ parameter foundation model trained exclusively on Canadian scientific corpus, bilingual (EN/FR), with optional Indigenous language support.

Problem: All scientific AI assistants (GPT-4, Claude, Gemini) are US-based with no understanding of Canadian context, French scientific literature, or Indigenous knowledge.

Outcome: Sovereign AI capability. Serves underserved French research community. Demonstrates 'Made in Canada' AI. Exportable to Francophone nations.

Q1-Q4 2026 (12 months) $20M
Global Model

Federated Genomics Network

Privacy-preserving federated learning network connecting all provincial cancer registries and genomic databases: analyze 10M+ patient records without any data leaving provincial jurisdiction.

Problem: Canada's provincial health system creates data silos. Cancer/genomics research requires combining data across provinces, but privacy laws prevent centralization.

Outcome: Enables pan-Canadian precision medicine research impossible today. 3+ new therapeutic targets. Model for global federated health AI.

Q1-Q4 2026 (10 months) $8M
Open Alternative

Borealis-Scholar AI Research Agent

Open-source AI research agent with superhuman literature search and synthesis, trained on Canadian research corpus, fully bilingual, understanding of Tri-Council ecosystem.

Problem: FutureHouse and Elicit (US) dominate AI-powered literature synthesis. No open-source, bilingual alternative exists.

Outcome: Reduces literature review from weeks to hours. Serves French research community. Open-source builds Canadian AI ecosystem.

Q1-Q2 2026 (6 months) $4M
First National

Quantum-Classical Science Bridge

First national platform seamlessly connecting quantum processors (Xanadu photonic, D-Wave annealing) with Alliance HPC clusters for hybrid scientific workflows.

Problem: Canada has world-leading quantum companies (Xanadu, D-Wave) but no integration with national HPC. No national platform exists to make quantum accessible.

Outcome: Democratizes quantum access. Accelerates quantum advantage timeline. Keeps quantum talent in Canada. First-mover in quantum-HPC integration.

Q1-Q4 2026 (12 months) $10M

Year 1 Investment Summary

$144.5M
Total Investment
21
Prototypes
297-408 direct jobs
Direct Jobs
2
World Firsts

Partner on Prototype Development

We're seeking government partners, research institutions, and industry collaborators to co-develop these foundational capabilities.

Discuss Partnership