Insights

Canada has long underinvested in business capital, but with a raft of fiscal incentives and early signs of AI-related investments accelerating, that tide may be turning.

Still, we don’t expect rising AI investments to add to Canada’s gross domestic product growth in the same way as in the United States—much of the new spending in Canada, especially on computer equipment, is imported instead of domestically produced. The bigger economic gain will likely come later, as wider adoption of the technology shows up in faster productivity growth.

Adopting technical advances at speed can be disruptive. Yet the reality is that producing more output per hour worked (productivity growth) will be increasingly crucial to offset the downward pressure that aging demographics and rising retirements are putting on the labour force and potential GDP growth.

AI business investments are heavily weighted to equipment purchases…

The wave of new infrastructure investments that drove a disproportionate share of U.S. business investment growth in the past years has shown early signs of spilling over to Canada, with major new data centre projects breaking ground across the country.

These investments can be broadly broken into three categories: spending on software, spending on non-residential structures including new power facilities and data centres, and spending on machinery and equipment such as semiconductors and data servers to adapt and fill those structures.

In the U.S. where AI investments are more advanced, Federal Reserve modelling has shown that the biggest contribution to business investment by far has come from computer equipment purchases, followed by software investment, with a smaller amount tied to structures investment including data centres.

…that are mostly imported in Canada

Replicating this exercise in Canada confirms that businesses are ramping up AI investment. Purchases of computer and peripheral equipment surged 28% in volumes over the first half of 2026 (year-over-year) according to the national accounts data, and are set to rise further in Q3.

The catch, however, is that most of those equipment purchases are imported rather than domestically produced. As of September, Canadian imports of computers and peripheral equipment were 75% above a year ago—led by a jump in high-value processing units used in data centres.

And because imports subtract from GDP, the net impact on growth has so far been relatively neutral. In fact, imports of computer equipment have modestly outpaced related investment spending in Canada, suggesting small inventory buildup that’s not captured in this add-up.

Looking at other capital expenditures, software investment has grown more modestly in Canada in recent years relative to the U.S., not showing as many AI signals. This is in line with the reality that the U.S. remains dominant when it comes to developing foundation models and GenAI software.

Finally, investment in AI structures is harder to track in Canada as there are no dedicated line items capturing spendings on data centres or power facilities. Monthly production data pointed to some momentum in non-residential building and engineering construction since the spring, that will still add up to a smaller gain in overall structures investment relative to machinery and equipment.

Bulk of AI-related payoff in Canada will show up in productivity gains later

However, the real longer-run value of investments in new technologies has rarely been limited to the direct contribution to GDP from new construction and equipment purchases at the build-out stage. The bulk of gains has always come later, when new technology unlocks capacity to drive future productivity gains that are critical to sustaining wage growth for workers.

AI investments will likely follow the same playbook. Canada’s productivity growth to-date remains lacklustre—GDP per hour worked grew just over 1% in the last year—up modestly from prior years, yet still indistinguishable from the tepid multi-decade trend pre-pandemic.

But wider AI adoption in the future, implemented safely with oversight, has the potential to enable the existing workforce across sectors to produce more output with the same or fewer hours worked. And those productivity gains will be welcome – if not critical– against a backdrop of aging demographics.

Today, more than one in three Canadians is above the age of 55. Immigration caps could loosen from current levels, but rising retirements will continue to subtract from working age population and limit growth in the economy. Productivity growth will be key to support potential GDP growth over the long run.

---

About the author:

Claire Fan is a Senior Economist at RBC. She focuses on macroeconomic analysis and is responsible for projecting key indicators including GDP, employment and inflation for Canada and the US. 

---