Read the Study

Reclaiming the Stack:
A Feasibility Study

Reclaiming the Stack:
A Feasibility Study

Title: Reclaiming the Stack

Subtitle: Co-operative AI Infrastructure for Canada’s Solidarity Economy: A Feasibility Study

This study, produced by CanTrust Hosting Co-operative and Hypha Worker Co-operative with funding from Co-operators’ Co-operative Development Program, examines whether a genuinely co-operative alternative is feasible: one that preserves data sovereignty, democratic governance, and environmental integrity without asking organisations to accept inferior tools or prohibitive costs.

Our report covers:

  • Market analysis,
  • Technical feasibility (including original research on energy consumption),
  • Risk assessment,
  • And financial viability.

What we found was that such an alternative is both technically achievable, and clearly needed.

To maintain resilience, the cooperative sector must build its own capacity to manage emerging technologies rather than relying on extractive external systems. This groundbreaking study moves beyond critique to offer a practical, collective approach to digital ownership. While focused on Canada, the report is applicable to all cooperative economies concerned about data sovereignty and the development of collectively owned and managed sustainable AI systems.”

Sonja NovkovicSonja Novkovic, Academic Director, International Centre for Co‑operative Management
Too long, Didn’t Read

Reclaim the Stack: An infographic

  • How much energy does AI use? image
  • Perception vs Reality of AI energy usage statistic
  • Hardware matters energy comparison of hardware and cost
  • Statistic on AI tool usage from survey

Click slider buttons to pause.

Appendix A

As part of this study, CanTrust and Hypha technical teams performed rigorous , low-level monitoring of AI energy usage measured in kilowatt-seconds (kW-sec). They tested a wide variety of hardware configurations. Download the raw data findings.

CanTrust and Hypha technical teams performed rigorous, low-level monitoring of AI energy usage measured in kilowatt-seconds (kW-sec). They tested a wide variety of hardware configurations.

Next Steps

Join us in Realizing Phase 3

Join us in Realizing Phase 3

If this resonates with your organisation and you want to be part of a co-operative solution. Sign up to our list and we will include you in our follow up discussions.