Submitted by Robson Inc. with DVNet, VIVO Media Arts Centre, CubeCommons, and BC + AI

Funding Request: $5,000,000


Reclaiming AI for Public Good

In a moment when most AI headlines are written in Silicon Valley boardrooms, we’re writing a different story—from the unceded lands of the Coast Salish Peoples, in server rooms above artist-run centres and code sprints held in shared kitchens.

Robson Inc., an Indigenous-owned BC-based SME, in partnership with DVNet, VIVO Media Arts Centre, CubeCommons, and BC + AI, submits this proposal to the AI Compute Fund to deploy Emily AI: an open, modular orchestration platform designed for community use, nonprofit deployment, and sovereign data stewardship.

We’re requesting $5 million to scale Emily AI’s development and to deploy it in real-world, public-interest environments—building community-controlled compute infrastructure, fine-tuning ethical models, and creating skilled jobs for equity-deserving communities.

This isn’t just about access to GPUs. It’s about who gets to write the future of AI in Canada.


The Problem: Centralized Power, Disconnected Communities

Two issues are at the heart of Canada’s AI legitimacy gap:

  1. Lack of sovereign, public-interest compute infrastructure, especially for nonprofits and cultural organizations.
  2. Exclusion of Indigenous, equity-deserving, and creative communities from shaping how AI tools are built, trained, and governed.

This initiative confronts both.


The Solution: Emily AI + Community Compute

Emily AI is Robson’s flagship orchestration layer for AI. It’s modular. LLM-agnostic. Context-aware. And designed from day one to be integrated into community-owned data commons with secure APIs, memory-layered inference, and blockchain-based governance (via DVNet). Think of it as a local brain for collective knowledge work—capable of powering grant portals, cultural archives, climate dashboards, or governance toolkits.

We’ll build on DVNet’s decentralized infrastructure to train and deploy Emily’s agents across three anchor sites: