AI Depends on More Than Models

AI infrastructure brings important tradeoffs, including power demand, water use, environmental impact, community concerns, and the need for measurable sustainability claims.

AI conversations often focus on models, applications, prompts, and productivity gains. Those topics matter, but they don’t tell the whole story.

Large-scale AI also depends on infrastructure: compute capacity, energy, cooling systems, network connectivity, data governance, and jurisdiction. As AI moves from experimentation into production, these infrastructure questions become more important.

That makes Canada’s recent work with TELUS on sovereign AI infrastructure a useful example of how AI strategy is changing. The announcement is not only about data centers or graphics processing units. It reflects a broader shift in AI strategy, where countries and organizations are thinking more carefully about where AI runs, where data resides, and who controls the systems behind it.

The Announcement in Context

The Government of Canada and TELUS are advancing work under the federal Enabling Large-Scale Sovereign AI Data Centres initiative. The proposed project would expand Canada’s sovereign compute capacity and support the domestic innovation ecosystem, including academia, industry, startups, public institutions, and government organizations.

TELUS has described the project as an expansion of its Sovereign AI Factory network, with facilities planned in British Columbia, including Kamloops and Vancouver. According to TELUS, the proposed cluster would eventually scale to more than 60,000 high-performance GPUs and 150 megawatts by 2032.

It’s also important to note what the federal announcement says clearly: no funding has yet been committed or distributed. The current work involves engagement with promising proposals and exploration of future collaborative opportunities.

That detail matters because infrastructure announcements can easily sound more final than they are. In this case, the project represents a significant direction of travel, but not every detail has been finalized.

What Sovereign AI Infrastructure Means

“Sovereign AI infrastructure” can sound abstract, but the core idea is straightforward.

It refers to AI compute infrastructure that operates within a country’s borders, under that country’s legal, regulatory, and governance framework. For Canada, that means Canadian organizations can access advanced AI infrastructure while keeping data, intellectual property, and sensitive workloads closer to home.

TELUS describes its Sovereign AI Factory as infrastructure designed for domestic data residency, Canadian operations and support, and secure AI development from prototype to production.

The practical value is not only technical. Sovereign infrastructure can support clearer governance. It helps organizations answer questions such as:

• Where is the data stored?
• Where is it processed?
• Who operates the infrastructure?
• Which laws apply?
• What controls exist around access, security, and oversight?

Those questions become more important as AI moves into regulated, sensitive, or strategically important areas.

Why This Matters for Data Analytics

For data analytics teams, this announcement points to a broader reality: AI readiness depends on more than having access to a model.

Organizations need reliable data pipelines, clear governance, secure environments, and infrastructure that can support real-world workloads. Without those foundations, AI remains fragile. It may work in a prototype but fail when scaled into production.

This is especially relevant in fields such as healthcare, finance, public services, research, and critical infrastructure, where organizations must manage privacy, security, compliance, and public trust.

Before an organization can use AI responsibly, it needs to understand where its data is, how it moves, who controls it, and what infrastructure supports it.

That’s a data analytics issue as much as it is an AI issue.

Compute Capacity Is Becoming Strategic

AI models require substantial compute resources, especially for training, fine-tuning, simulation, and high-volume inference. Access to that compute can shape who gets to participate in advanced AI development.

If compute capacity is scarce, expensive, or located entirely outside the country, smaller businesses, academic researchers, public institutions, and startups may face barriers. They may also need to send sensitive data or intellectual property into environments governed by other jurisdictions.

That is why sovereign compute capacity has become part of national AI strategy. It’s not only about performance. It’s about access, control, resilience, and economic participation.

Minister Evan Solomon summarized the infrastructure side of AI directly in a LinkedIn post, noting that AI runs on “compute, energy, networks and talent.”

That framing is useful because it pushes the conversation beyond the visible layer of AI. The model is only one part of the system.

The Sustainability Question

AI infrastructure also brings tradeoffs.

Large data centers require power, cooling, land, network capacity, and local infrastructure planning. TELUS has emphasized the sustainability features of the proposed B.C. cluster, including clean energy, liquid cooling, reduced water use, and waste heat recovery that could help heat homes in Vancouver.

At the same time, CBC reported concerns from critics who argue that data center development needs stronger environmental review, community consultation, and regulation. Those concerns include power demand, water use, and the risk of moving too quickly before local impacts are fully understood.

Both sides of that conversation matter.

AI infrastructure may support innovation and economic growth, but it should still face clear questions about resource use, environmental impact, transparency, and public accountability. Sustainability claims need measurable standards, not just broad statements.

What Organizations Should Ask

For organizations considering AI services, infrastructure, or vendor partnerships, this announcement highlights several practical questions.

Data and jurisdiction

• Where will data be stored and processed?
• Which laws apply?
• Can the organization clearly explain the data flow?

Governance

• Who operates the infrastructure? Who has access?
• What controls exist for privacy, security, monitoring, and auditability?

Production readiness

• Can the environment support production workloads, or is it mainly useful for experimentation? How does it handle scaling, reliability, and integration with existing systems?

Sustainability

• What energy sources power the infrastructure? How much water does it use? How are cooling efficiency and waste heat recovery measured?

Strategic fit

• Does the infrastructure support the organization’s actual use cases, or is it being adopted because AI is receiving attention?

These questions are not only for technology teams. Legal, compliance, procurement, sustainability, business operations, and data governance teams all have a role to play.

AI Strategy Is Becoming Infrastructure Strategy

The key point is not that every organization needs to build its own AI infrastructure. Most won’t.

The larger point is that AI strategy now reaches beyond model selection, prompt design, and application development. It includes infrastructure, governance, sustainability, data residency, and jurisdiction.

Canada’s work with TELUS offers a useful example of that shift. It shows how AI readiness increasingly depends on the underlying systems: the compute, networks, energy, cooling, and governance structures that determine how AI gets built and deployed.

For data analytics professionals, this is an important development to watch.

AI may appear to users as software, but at scale it depends on physical infrastructure and disciplined data governance. Organizations that understand that connection will make better decisions about adoption, risk, compliance, and long-term value.

About TELUS

TELUS is a global communications technology company headquartered in Vancouver, Canada. The company provides broadband, wireless, digital, health, agriculture, and consumer goods technology services across more than 45 countries. TELUS is focused on using technology to improve customer experiences, strengthen communities, and support better health and social outcomes.

Disclosures

I am an employee of TELUS Health, which is part of TELUS. This post reflects my own perspective and is based on publicly available information.

Sources

Government of Canada - Government of Canada and TELUS advance work to build sovereign AI infrastructure
TELUS - TELUS and Government of Canada advance work to scale Canada’s sovereign AI infrastructure
TELUS - Sovereign AI Factory
LinkedIn Post - Evan Solomon, Canada’s Minister of AI and Digital Innovation
CBC News - Plan unveiled for 'sovereign AI data centre' cluster in Kamloops, Vancouver