“Canada has what the world wants,” Canadian Prime Minister Mark Carney told his Davos audience in January 2026 — a message he repeated at September’s Canada Investment Summit. His vision of principled pragmatism aims to balance upholding Canadian values while working with partners who do not necessarily share them. Still, as Ottawa broadens these partnerships, it must also understand what it is agreeing to in terms of access and technology transfers, and how it plans to verify compliance and enforce the terms. Canada’s nuclear relationship with India offers a cautionary lesson on the need for clear restrictions and effective oversight.
The case for diversification and a stronger commercial perspective is compelling. US tariffs have sharpened Canada’s need for other export markets and investment sources. As The Economist notes, in 2025, more than 70 percent of exported goods went to the United States and investment in Canada from 2014 to 2025 rose by only 4.2 percent, compared with the 51 percent increase in the United States. Clearly, more diversified exports and new investments are necessary. September’s investment summit in Toronto was part of that effort to attract capital, and the EU–Canada summit in Montreal is the next test. As capital flows in, Ottawa needs to be clear and transparent about the terms on which it comes. Alongside the expected economic benefits, Canadians should know who will control strategic assets, what access foreign partners will gain to Canadian technology and expertise, and what recourse Canada retains if commitments are broken.
While EU member states are like-minded democracies, Carney’s goal of doubling exports to markets outside the United States over the next decade to some CDN$600 billion annually will broaden the range of countries that Canada exports goods and technologies to and raise new questions. Canada is, for example, exploring the sale of AtkinsRéalis’s CANDU nuclear reactors to Turkey. Such agreements raise a number of safety, financial and strategic concerns given regional security dynamics and Turkey’s economic and political stability. Canada’s nuclear relationship with India offers a historical lesson in why these questions belong at the negotiating table and require public scrutiny from the outset.
A Lesson from India
In 1956, Canada agreed to supply India with a research reactor subject to a commitment to peaceful use. The meaning of “peaceful use” was never clearly defined, although there was an assumption of good faith. But in May 1974, India conducted its first nuclear test using plutonium produced by the Canadian-supplied reactor and justified it as a “peaceful” nuclear explosion.
The weakness went beyond the ambiguous wording of “peaceful purposes” in the original agreement. The 1956 pledge came with no requirement for inspections or controls that would have allowed Canada to verify India’s commitments. Later, when Canadian inspectors did visit, their checks covered only Canadian-supplied fuel. As the head of Canada’s Atomic Energy Control Board later conceded, these checks were made futile for detecting weapons development. By June 1968, Canadian inspectors had encountered information suggesting India might be developing a nuclear device. That warning did not prevent the test six years later and, in the end, Canada had helped transfer a critical capability without securing effective controls over its use.
The lesson in contemporary dealmaking for Canada today is that restrictions need both a clear meaning and a credible means of verification. The 2010 Canada–India nuclear agreement explicitly requires peaceful and non-explosive use, provides for International Atomic Energy Agency safeguards on covered nuclear materials and equipment, and requires written consent for transfers to third countries. These provisions establish obligations and verification requirements — and although they do not guarantee compliance or eliminate every strategic risk, they do help minimize vulnerabilities.
Still, those provisions matter for Cameco’s uranium supply agreement with India, announced in March 2026. The company estimates the nine-year contract’s value at CDN$2.6 billion, with deliveries expected to begin in 2027 and a stated purpose of supplying India’s civilian nuclear reactors.
However, safeguards governing imported uranium do not settle a separate question. That is, whether those imports allow India to devote more domestic uranium to military purposes — a concern that arms-control researcher Ernie Regehr raised in 2009. This question warrants a current assessment, rather than assuming safeguards can resolve the issue.
In that sense, the distinction between a declared purpose and an enforceable restriction should inform Canada’s future agreements on sensitive sectors, such as critical minerals, defence technology and artificial intelligence (AI).
Ottawa has already started to consider types of minerals that require protection and has identified 34 minerals and metals that it deems essential to Canada’s economic and national security. Geopolitical challenges have created a wider awareness of dependence on a single buyer or processor that Canada is likely already considering in the deals it negotiates. For defence exports and dual-use technologies, Canada should strengthen its efforts on countering “diversion” — that is, retransfer to a third party, whether intentional or not — and how compliance will be monitored. Given that not all exports are government to government, a better understanding of where companies do business is important for assessing downstream risk. For AI partnerships, access to data on Canadian citizens and companies is worth examining.
The controls will inevitably differ, as nuclear safeguards —while applicable in some cases — cannot simply be transplanted to software or other technologies. Canada should then be clear on three key questions: What is being transferred or given access to; how can restrictions be verified; and what can it do if they are breached?
Where continuing supply, servicing or technical support gives Canada influence, agreements should also specify the terms for suspending it. Some risks will justify narrower cooperation, while others may justify declining a deal outright even if the short-term gains may look appealing. Those tests should apply to established allies, including the European Union, as well as prospective partners.
Similarly, Ottawa should clarify the principal restrictions governing strategic agreements, name compliance overseers and consider what would trigger a review. Doing so requires technical specialists, diplomats and regulators with the resources to scrutinize deals and monitor them long after agreements are signed. These capable individuals would need to be recruited or trained and would also need increased resources. Cuts to Global Affairs Canada show a disproportionate impact on positions abroad — an unfortunate trend given current geopolitical challenges and stated aims of diversifying exports and the need to monitor implementation of agreements.
In 1956, a promise of peaceful use was enough, but it took 18 years and a nuclear test to show what that promise was worth. As Canada pursues a wider range of strategic partners, including governments that do not share Canadian values, putting principled pragmatism into practice requires thoughtful analysis of risk and trade-offs. Canada has what the world wants, and its leverage is greatest before a deal is signed. Cutting the capacity to negotiate and enforce an agreement’s terms risks leaving others to set them.
