Canada has decided to formalize—on a strict timetable—one of the most delicate pieces in the global race for Artificial Intelligence: where the computing infrastructure needed by companies, universities, and labs will be hosted to compete without relying entirely on external infrastructure. The Ministry of Innovation, Science, and Economic Development (ISED) has launched a proposal submission process to enable large-scale commercial AI data centers, with a threshold that reflects the type of project Ottawa is seeking: more than 100 megawatts (MW) of planned total capacity.
The proposal deadline is precisely defined: from January 15 to February 15, 2026. The window is short, and the message is clear: the government wants to identify projects with real potential—not preliminary ideas—and accelerate discussions with the private sector through memorandums of understanding (MOUs) with the selected proponents.
The goal: “sovereignty” and ecosystem-wide capacity
In the language of the call, the key word is sovereign. ISED envisions these centers as infrastructure intended for a “broad spectrum” of Canadian clients, and positions this effort within a broader strategy: to increase domestic computing capacity so that the innovation ecosystem—in academia and industry—does not depend on foreign decisions, cloud bottlenecks, or access to accelerators.
This approach is based on an assessment already outlined in Budget 2025: the adoption of digital technologies and the rise of AI have created a critical need for processing power for services and products. In that same fiscal document, Canada proposes CA$925.6 million over five years (from 2025-26) for a sovereign public AI infrastructure aimed at increasing computing availability and access for public and private research. Of this amount, CA$800 million would come from funds already allocated within the fiscal framework. Additionally, the budget indicates that the Minister of Artificial Intelligence and Digital Innovation will collaborate with industry to identify promising projects and formalize them through MOUs, and it opens the possibility for the Canada Infrastructure Bank to invest in AI infrastructure initiatives.
Meanwhile, ISED’s call seeks projects that not only grow capacity but do so under national conditions: Indigenous participation, minimized environmental impact, and a “maximized” use of Canadian partners and supply chains. In a sector where servers, networks, chips, and critical components often depend on global supply chains, this preference introduces a significant tension between industrial ambition and operational realism.
Over 100 MW: a threshold that reshapes the conversation
The bar of 100 MW is not a technical detail but a statement of scale. It’s not about expanding IT “rooms,” but about designing facilities capable of handling intensive model training and inference loads, as well as associated services (mass storage, high-performance networks, advanced cooling). Canadian media have noted that, while a conventional data center operates within much lower ranges, AI-focused projects tend to require orders of magnitude higher, both in computing density and in energy and cooling needs.
This aspect explains why the call emphasizes energy considerations: promising MW capacity is not enough; proponents must detail how it will be achieved, what efficiencies will be applied, and how environmental footprints will be reduced. In 2026, energy is the bottleneck separating mere promises from actual, buildable projects.
What Ottawa assesses: criteria, governance, and readiness
ISED requests that proposals explain how they align with specific criteria. The list is broad, essentially covering the “checklist” for a megacenter:
- Economic benefit and ecosystem contribution (impact on innovation, employment, services for businesses and research).
- Indigenous participation (from formal collaboration to deeper participation structures).
- Sovereignty factors (control, location, critical dependencies).
- Energy (supply, planning, sustainability).
- Performance capabilities (types of loads supported and guarantees).
- Level of readiness (project maturity and path to execution).
- Cost and structure (funding, partners, development model).
The process also extends beyond ISED: by submitting the form, the proposer agrees that the ministry may share the proposal with federal partners (such as investment agencies, funding bodies, or the major projects office), as well as with provincial or municipal departments and agencies if necessary to evaluate or refer the project. The information, according to the text, will be handled in line with Canadian access to information and privacy laws.
Another important nuance: the instrument announced is MOUs, not construction contracts or public procurement awards. Media coverage from the sector reports that government spokespersons emphasize these MOUs as mechanisms to facilitate dialogues, without implying automatic funding or binding commitments at this stage.
A defining moment: the government will use AI to review proposals
In a move that blends efficiency with symbolism, ISED has announced it will use AI to help summarize the proposals received. The implicit message is twofold: on one hand, the agency expects volume and complexity in the submissions; on the other, it normalizes the use of AI as an internal management tool—just as the process is designed to deploy AI infrastructure.
How to submit a proposal (no fuss)
The process is straightforward and intentionally direct: Canadian companies or consortia that meet the requirements must request the form and send the completed document via email before February 15, 2026. ISED manages the contact channel and form submission at the address [email protected].
A sign of industrial strategy, more than an isolated announcement
The underlying message is that Canada is shifting from a conversation about AI—models, talent, adoption—to one about critical infrastructure: energy, land, cooling, networks, operation, and control. In an international landscape where processing capacity has become a competitive advantage, ISED’s call functions as a thermometer: the country seeks hyperindustrial-scale projects, but within an explicit framework of sovereignty, participation, and sustainability.
Frequently Asked Questions
What exactly does “sovereign AI data center” mean in Canada’s call?
It refers to infrastructure designed to serve Canadian clients with sovereignty criteria—control, governance, and critical dependencies—and aimed at increasing domestic computing capacity for research and the private sector.
Why is the minimum threshold 100 MW, and what does this mean technically?
Because this level targets facilities prepared for large-scale AI workloads (high GPU density, intensive networks, and storage). It involves significant energy, cooling, permitting, and supply chain challenges far beyond those of a traditional data center.
Which criteria weigh most heavily in ISED’s proposal selection?
Beyond economic impact, Ottawa prioritizes a clear path to implementation, Indigenous participation, environmental mitigation, energy considerations, sovereignty factors, and the use of Canadian partners and supplies.
How do I submit a proposal, and what is the deadline?
The form must be requested and emailed to [email protected]. The deadline to submit is February 15, 2026.
Source: ised-isde.canada.ca