Build Canadian Sovereign AI Models
We design and develop AI models aligned with Canadian priorities, industries, laws, and public-interest objectives.
Accessible, responsible, Canadian-governed AI
Canadian Sovereign AI, Built for Public Benefit.
GoAutomate AI Institute is a not-for-profit organization focused on advancing Canadian sovereign artificial intelligence through AI models, tools, and resources designed for Canadian organizations.
Our Purpose
AI should broaden opportunity and access to capability.
Canadian sovereign AI must be intentionally designed, responsibly governed, and made accessible to the organizations and communities that will shape Canada’s future.
AI should enhance human ingenuity and extend our capabilities.
Our work is rooted in the belief that artificial intelligence should serve people, strengthen organizations, and distribute benefits widely.
Transformative technologies can concentrate or broaden opportunity. They can make life easier for a few, or they can expand opportunity for many.
GoAutomate AI Institute exists to ensure that artificial intelligence strengthens Canadian organizations, supports public benefit, and contributes to a more resilient national AI ecosystem.
AI must reflect Canadian values, Canadian ethics, Canadian governance, and Canadian interests.
What We Do
Building practical AI capabilities for Canadian organizations.
We focus on models, tools, education, and governance frameworks that help organizations adopt AI responsibly and effectively.
Make AI Accessible
We make AI models, tools, documentation, and educational resources available to organizations across Canada.
Support Responsible Adoption
We help organizations adopt AI with ethics, governance, transparency, security, and accountability.
Strengthen AI Independence
We support sovereign capabilities that reduce dependency on foreign-controlled AI models and infrastructure.
Why Sovereign AI Matters
Canada needs AI systems that reflect Canadian priorities.
Canadian organizations need AI to compete globally, improve productivity, and modernize how they operate. However, Canada cannot rely solely on AI systems controlled, governed, and optimized outside of the country.
A Canadian sovereign AI ecosystem can provide a more secure, resilient, and accountable path forward.
- ✓Reduce dependency on foreign AI infrastructure.
- ✓Support Canadian laws, policies, and governance standards.
- ✓Reflect Canadian values and social priorities.
- ✓Improve national security and economic resilience.
- ✓Enable broader access to AI across public and private sectors.
- ✓Help ensure that the benefits of AI are shared widely.
Our Goals
A practical path toward Canadian sovereign AI.
Canada does not need to participate only in the race toward larger frontier models. Purpose-built, efficient, Canadian-governed models can deliver meaningful outcomes in practical environments.
Accelerate Canadian AI Adoption
Support organizations in adopting AI to improve productivity, reduce manual work, and unlock operational value.
Build Efficient Models
Create sophisticated, purpose-built models that can operate in smaller, more cost-effective environments.
Promote Human-AI Collaboration
Keep people at the helm while allowing AI to enhance decision-making, creativity, quality, and speed.
Advance Responsible Governance
Promote AI systems that are steerable, explainable, accountable, secure, and aligned with Canadian expectations.
Share Benefits Broadly
Ensure AI-driven productivity contributes to shared prosperity, not only to a select group of technology providers.
Protect Strategic Independence
Reduce reliance on foreign models and infrastructure while advancing Canadian innovation and leadership.
Research & Open Models
Our first release: the Terna model family.
Terna — from the Latin terni, “three each” — is our family of ternary‑weight language models: every weight constrained to just three values, { −1, 0, +1 }. It is our bet on the next wave of efficient AI — capability that scales down in cost as readily as it scales up in ability.
≈8–10× smaller
At ~1.6 bits per weight, ternary models shrink roughly an order of magnitude versus half precision — and about half the size of 4‑bit quantization.
Multiply‑free
A weight in { −1, 0, +1 } turns the costly multiply into add, subtract, or skip — with roughly a third of the work vanishing as structured sparsity.
Capability preserved
We reach ternary through distillation — a student trained to match a high‑precision teacher — so the constraint is learned, not imposed. Subtract cost, not intelligence.
Runs broadly
The small footprint fits capable models onto edge devices and revives legacy datacenter GPUs, while giving modern accelerators far more concurrent users.
Terna‑E2B — open weights, pre‑release.
A ternary distillation of Gemma‑4‑E2B — a capability‑dense model with a transparent, Western open‑weight lineage — published as GGUF and runnable on llama.cpp. This is an early (~1B‑token) checkpoint shared for community evaluation; production weights (~15B tokens) with full benchmarks will follow. The mathematics and methodology are documented in our first technical report.
Our Vision
A Canada where every organization can responsibly access and use advanced AI.
We envision hospitals, health authorities, public institutions, businesses, researchers, and not-for-profit organizations using AI to become more efficient, more innovative, and more resilient.
Help build Canada’s sovereign AI future.
GoAutomate AI Institute welcomes collaboration with Canadian organizations, healthcare systems, public institutions, researchers, policymakers, and technology leaders who share our commitment to responsible and accessible AI.