Ottawa Spent More Than $800 Million on AI in Just Three Years, New Data Show

35,000+ smart investors are already getting financial news, market signals, and macro shifts in the economy that could impact their money next with our FREE weekly newsletter. Get ahead of what the crowd finds out too late. Click Here to Subscribe for FREE.

Ottawa’s rush into artificial intelligence is no longer a future-tense policy debate. New federal data show more than $800 million in AI-related contracts, subscriptions, licences and arrangements since 2023, pulling everything from payroll modernization to defence mapping, tax analytics and veterans’ disability claims into the same national conversation.

The spending total is not a simple receipt for “AI tools.” It blends major modernization projects, support for Canadian AI firms, cloud-enabled software, small subscriptions and departmental pilots. Still, the figure lands at a sensitive moment: Ottawa is promising a more productive public service while Canadians are asking how much AI will cost, who benefits, and whether enough safeguards are in place.

The $831 Million Figure Is Likely Only Part of the Story

The new total comes from data requested by Conservative MP Jagsharan Singh Mahal, who asked federal departments, agencies and Crown corporations to disclose AI-related contracts, subscriptions, licences, pilot projects, memoranda of understanding and other arrangements. The request covered the period from Jan. 1, 2023, to March 9, 2026. After compiling the responses, The Canadian Press calculated the disclosed total at roughly $831 million.

That number should be treated as a floor, not a ceiling. Some organizations did not fully comply, and others said they could not provide centralized figures. The Communications Security Establishment and Canadian Security Intelligence Service declined to share details beyond public information, while the RCMP said the information was not available in one centralized database. That means the visible spending already tells a major story, but the complete picture is still partly hidden inside Ottawa’s own record-keeping gaps.

Two Big Deals Dominate the Spending Total

Two previously announced items account for a large share of the disclosed total: a $350 million public-service contract with Dayforce tied to replacing the troubled Phoenix pay system, and a federal investment of up to $240 million in Cohere, the Toronto-based AI company. Together, those two commitments represent nearly three-quarters of the roughly $831 million compiled from departments and agencies that responded.

That matters because it changes how the headline should be understood. A spokesperson for AI Minister Evan Solomon said the $800 million figure should not be read as one single total for government purchases of AI tools. It combines very different categories, including AI-enabled software, cloud services, wider digital modernization projects where AI may be one component, and programs supporting Canadian AI companies and compute capacity. In other words, Ottawa is not just buying chatbots; it is trying to rebuild pieces of government technology around AI-era infrastructure.

Phoenix’s Long Shadow Makes Dayforce a High-Stakes Bet

The Dayforce contract is especially sensitive because it sits in the shadow of Phoenix, the federal payroll system that became synonymous with public-service pay errors. Ottawa says Dayforce is part of a broader HR and pay transformation meant to replace Phoenix and modernize systems used across government. The current federal pay system delivers biweekly pay to an average of 431,000 current and former employees, with roughly 13.4 million payments worth about $40.1 billion in 2024.

The Auditor General has already warned that the transition carries serious risks. In 2024–25, the federal pay system processed $38 billion in pay for more than 430,000 current and former public servants. The preliminary estimate for the new HR and pay system was more than $4.2 billion, excluding some transition costs. Ottawa shortened the schedule to move departments and agencies to Dayforce by about three years, partly to reduce the cost and complexity of running two pay systems at once. That makes execution just as important as procurement.

Cohere Shows Ottawa’s Sovereign AI Strategy in Action

The Cohere funding is different from the Dayforce contract. It is not simply a departmental software purchase; it is part of Ottawa’s effort to build domestic AI compute capacity and support a Canadian-born AI firm competing globally. The federal government finalized an investment of up to $240 million in Cohere’s $725 million project to bring more domestic compute capacity to Canada and support the development and scaling of AI capabilities.

Cohere also signed a memorandum of understanding with the federal government in 2025 to explore ways AI could improve public-service operations while supporting Canadian commercial AI capabilities. The company has positioned itself around enterprise and government-grade AI, including customized tools for regulated sectors. In 2026, Reuters reported that Cohere agreed to acquire Germany’s Aleph Alpha as part of a push into secure, sovereign AI solutions for government and business customers. For Ottawa, Cohere has become both an economic-development play and a test case for buying Canadian AI.

Defence, Tax and Veterans Files Show AI Moving Into Core Government Work

Beyond the two largest deals, the disclosed spending shows AI spreading into areas that directly touch public administration. Public Services and Procurement Canada and Innovation Canada led the spending because of Dayforce and Cohere, followed by National Defence at $83.7 million, the Canada Revenue Agency at $29.9 million, and Veterans Affairs Canada at $25.1 million. National Defence also had the highest number of contracts, with 104 listed entries.

The examples are concrete. The CRA signed a $17.5 million contract with SailPoint Technologies for machine-learning analytics used for applications such as automatically flagging high-risk users. National Defence spending included a $6.3 million deal with Ecopia for mapping software that identifies land features and objects from satellite imagery. Veterans Affairs said nearly all of its spending went toward an AI tool that processes veterans’ medical records and generates claim summaries to reduce document volume for disability adjudicators. These are not abstract experiments; they involve taxes, defence mapping and veterans’ files.

Small Subscriptions Reveal a Bottom-Up AI Rush

The data also included hundreds of smaller entries, ranging from a few hundred dollars for a ChatGPT subscription to multimillion-dollar contracts with major vendors. Some examples were everyday productivity tools, such as AI captioning and transcription software. Others were bigger enterprise arrangements, including a $12.1 million agreement between the Bank of Canada and Microsoft to implement M365 Copilot.

This is where the story becomes more human. Inside government offices, AI is not always arriving as a grand strategy. Sometimes it looks like a team buying transcription help for meetings, a unit testing automated captioning, or analysts experimenting with tools that summarize and sort large volumes of text. That bottom-up adoption can make work faster, but it also creates governance challenges. A few hundred-dollar subscription may feel harmless, yet the real questions are what data goes into the tool, who reviews the output, and whether departments are tracking usage consistently.

Ottawa’s AI Spending Mirrors a Bigger Business Trend

The federal government is not operating in isolation. Statistics Canada reported that 12.2% of Canadian firms used AI to produce goods or deliver services in 2025, double the 6.1% share reported in the second quarter of 2024. Another 14.5% planned to adopt AI within the next 12 months. Adoption was highest in information and cultural industries, professional, scientific and technical services, and finance and insurance.

The same data suggest that AI is changing workflows more quickly than headcounts. Among businesses that had used AI, 89.4% reported no change to employment levels in 2025. The most common changes were developing new workflows, training current staff, purchasing cloud services, and adjusting data practices. That helps explain why Ottawa is interested: AI may be less about replacing entire jobs immediately and more about changing how files, claims, maps, payments and service requests move through large organizations.

Productivity Is the Promise, But Proof Is Still Early

The policy argument for federal AI spending is straightforward: Canada has weak productivity growth, government services are often slow, and AI could help departments process information faster. The Bank of Canada has also said there is no evidence so far of widespread worker displacement from AI, while noting that the technology is more likely to transform tasks than eliminate jobs outright. It has seen early evidence of small productivity gains, but the broader impact remains uncertain.

That uncertainty is important. A tool that summarizes records for an adjudicator, flags unusual account access, or maps satellite imagery can create real efficiencies if it is accurate, secure and well supervised. But AI does not automatically fix messy data, outdated processes or poor management. In some cases, it can simply make old problems move faster. The public-service challenge is not only to buy AI, but to prove that the tools reduce backlogs, improve service quality, protect privacy and save money over time.

Transparency Is Becoming as Important as Technology

The spending data exposed a basic accountability problem: Ottawa does not appear to have one clean, centralized view of all AI-related spending. Some departments reported detailed information, while others could not provide comparable data. That makes it harder for Parliament, journalists and taxpayers to separate small experiments from major procurement decisions, or software subscriptions from strategic industrial investments.

Federal guidance already recognizes the risks. Treasury Board guidance tells federal institutions to be cautious with generative AI, evaluate risks before use, and involve legal, privacy, security and other stakeholders before deploying tools for public use or service delivery. The Directive on Automated Decision-Making also requires departments to assess impacts, be transparent, ensure quality, provide recourse and report publicly when automated systems support administrative decisions. The challenge now is whether those principles can keep up with the speed and variety of spending.

The Governance Race Is Now Catching Up With the Spending

Ottawa has created the Canadian Artificial Intelligence Safety Institute to advance AI safety research and work with international partners on risks such as synthetic content, impersonation, fraud and advanced systems that may hinder human oversight. The government has also promoted voluntary codes, public-sector guidance and compute investments as part of a wider AI agenda. Those moves show that AI policy is no longer limited to innovation funding; it is also about public trust.

Still, Canada’s attempt at a comprehensive AI law has been uneven. The proposed Artificial Intelligence and Data Act was tied to Bill C-27, which did not become law after Parliament was prorogued in early 2025. That leaves Ottawa leaning heavily on guidance, directives, procurement rules and department-level controls while spending continues. The $831 million figure is therefore more than a budget story. It is a test of whether the federal government can move quickly enough to modernize services without outrunning transparency, privacy and accountability.

This Options Discord Chat is The Real Deal

While the internet is scoured with trading chat rooms, many of which even charge upwards of thousands of dollars to join, this smaller options trading discord chatroom is the real deal and actually providing valuable trade setups, education, and community without the noise and spam of the larger more expensive rooms. With a incredibly low-cost monthly fee, Options Trading Club (click here to see their reviews) requires an application to join ensuring that every member is dedicated and serious about taking their trading to the next level. If you are looking for a change in your trading strategies, then click here to apply for a membership.