RC122 JanFeb 2026 - Magazine - Page 30
LEGAL
BEYOND CODE
Powering AI; Canada’s Moment?
by David Tournier and Jessy Ménard
HE RISE OF ARTIFICIAL INTELLIGENCE (AI) is no longer just
about algorithms, data, and software. Once con昀椀ned to the minds of computer scientists, AI
now depends on massive physical infrastructure:
expanded electricity generation, resilient transmission and distribution grids, and high-performance
data centres, supported by advanced semiconductors,
critical minerals, and complex global supply chains. As AI
penetrates every sector, from defense to 昀椀nance, it is driving surging demand for reliable, a昀昀ordable, low-carbon
power; leading-edge hardware manufacturing; and the
construction and modernization of digital infrastructure.
This shift creates material challenges as well as opportunities, particularly for resource rich countries such
as Canada. This article 昀椀rst outlines AI’s infrastructure
requirements and then examines how Canada could position itself as a key player by leveraging its energy and
critical mineral endowments, while identifying critical
blind spots in current strategy.
T
David Tournier is a
partner and head of
Lavery’s Infrastructure
and Projects Group.
Jessy Ménard is an
associate at Lavery.
Infrastructures Required for AI
THE NEED FOR STABLE ENERGY
The electricity demand driven by the expansion of data
centres and AI applications is surging. The International
Energy Agency (IEA) estimates that data centres consumed approximately 415 terawatt-hours (TWh) in 2024,
roughly 1.5 per cent of global electricity consumption. This
昀椀gure could more than double by 2030, to roughly 945
TWh, with AI a primary growth driver. In some economies, data centres could represent more than 20 per cent
of electricity demand growth by 2030, while in the United
States, demand could grow more than 30x by 2035, to 123
gigawatts (GW), from four GW in 2024.
Meeting this rising demand requires new generation
capacity. Building renewable, nuclear, or gas-昀椀red plants
poses signi昀椀cant regulatory, environmental, and logistical
challenges. Renewables face land-use constraints, and wind
and solar intermittency is at odds with AI’s round-the-clock
power requirements. Nuclear entails high upfront costs,
licensing and safety requirements, and long lead times.
Several recent data centre and AI infrastructure projects,
including those led by xAI, Oracle, and Meta have turned to
natural gas generation, citing its speed of deployment and
30—RENEW CANADA –JANUARY/FEBRUARY 2026
ability to bypass grid interconnection delays.
The IEA emphasizes that accelerating deployment of
clean generation is essential to meet AI-driven demand
and climate commitments. In practice, countries must
balance the urgency of capacity expansion with decarbonization targets, an increasingly complex policy
challenge.
The stability and scalability of AI infrastructure also
depend on grid modernization and the strategic siting of
data centres. The IEA emphasizes that “a sole focus on
increasing electricity generation won’t be enough […]
countries must also think about their infrastructure.” Deloitte similarly notes that “the AI ambitions of the [U.S.]
government and industry come up against the grid’s
capacity to power or even interconnect data centers, as
there is currently a seven-year wait for some requests to
connect to the grid.”
MATERIAL NEEDS FOR AI INFRASTRUCTURE
Beyond energy, AI relies on a wide range of materials:
to build data centres (concrete, steel, copper, cooling
systems) and to equip them (semiconductors and specialized chips, cabling, rare earths, and high-purity metals).
RENEWCANADA.NET
