In 2026, Canada’s federal regulator publishes the first comprehensive assessment of the law requiring digital giants to pay news media. As of August 19, 2026, $102,544,410 had been distributed to 425 beneficiaries; year one had also distributed $96,732,819 to 445 beneficiaries, and Google is the only platform that participated. Meta refused to comply. Financing local press without restoring its audience raises a scope question that this assessment makes visible.
The Essential Points
- According to the independent auditor’s report published by the CRTC, $102,544,410 had been distributed to 425 beneficiaries as of August 19, 2026, for year 2.
- Google is the only major platform to have concluded the agreement provided for by law; Meta blocked access to news links in Canada rather than negotiate.
- The mechanism primarily finances news companies without directly regulating recommendation algorithms; Google remains a search intermediary for news, while Meta blocked the availability of news content on Facebook and Instagram in Canada.
- Canadian law is being closely watched in Europe and Australia as a possible model, but its first assessment invites us to distinguish between financial compensation and editorial sovereignty.
Google Alone at the Table, Meta at the Door
The Online News Act, adopted in 2023, required major digital platforms to negotiate a commercial agreement with Canadian media or submit to binding arbitration. The idea was straightforward: if a platform derives value from journalistic content by distributing it to millions of users, it must share part of that value with those who produce it.
Google chose to negotiate. The agreement reached provides for an annual contribution of $100 million Canadian, paid into an independently administered fund. Meta did the math differently. Rather than pay, the group cut access to Canadian news links on Facebook and Instagram starting in August 2023. Canadians using these two networks as a news source found themselves without press content in their feeds.
Meta believed that journalistic content was not essential to it.
This choice reveals something important about the nature of the power dynamic. For a platform whose model rests on engagement and retention, news content represents a fraction of what keeps users. Removing it costs less than paying for it. Google, whose search engine directly values informational content, calculated otherwise.
In practice, the law applies to only one of the two major platforms in question. This is a structural limit of any approach that depends on platforms’ willingness to participate, not a failure of enforcement.
425 Newsrooms, Considerable Disparities
The $102 million distributed reached 425 beneficiaries, according to the CRTC report. Distribution follows a logic of audience and reach: large national press groups receive higher amounts than local weeklies. For a regional media outlet whose annual budget is measured in hundreds of thousands of dollars, a payment of tens of thousands represents a significant contribution. For a national network, the same sum is marginal.
This mechanism reproduces a tension already present in traditional advertising models: media that need it least capture a larger share. The question of weighting—that is, how funds are distributed between large and small, between national and local, between print and radio—has structured debate since the law’s adoption.
The CRTC did not publish the complete breakdown by beneficiary in its audit report, but regional press professional associations reported that amounts received by small newsrooms rarely cover more than one or two full-time positions. For a newsroom that has lost five journalists in ten years due to lack of advertising revenue, this is a floor, not a reconstruction.
The Limits of Funding Without Audience
Circulation results from the objectives and signals inherent to each system, but also from public interest, competition, and editorial adjustments such as local prioritization. Recommendation systems often use engagement signals, but their objectives and parameters can also include relevance, satisfaction, quality, diversity, and other constraints. The absence of a direct territorial objective varies by platform and their configurations. Content covering a school board meeting in a small municipality may generate lower engagement than other content, reducing its chances of algorithmic recommendation, even if it were produced to the same professional standards as a nationally oriented article. The mechanism distributes compensation based on full-time positions allocated to producing original online news, under a law regulating digital news intermediaries.
Local press’s economic model fractured over two decades. Advertising revenue migrated to platforms, which offer advertisers targeting that local newspapers cannot match. The direct consequence was a reduction in newsroom teams, closure of titles, and the emergence of what American researchers call “news deserts”: entire geographic areas without local journalistic coverage.
The law organizes compensation without directly regulating algorithms, while potentially indirectly supporting certain distribution and promotion activities. An article produced by a small rural newsroom in Ontario receives indirect support through the Google fund, but the law does not directly regulate algorithms and a media outlet’s visibility does not depend entirely on them.
This is the central gap that the 2026 assessment brings to light. Intermediate formats are losing the algorithmic race: local media, which by nature produce niche content, are structurally disadvantaged in optimization systems designed for large-scale engagement. Funding preserves or strengthens certain production capacities; its effect on the place of local media in public space is not settled by the audit.
The distinction between financial survival and audience sovereignty is the crux of the problem. A newsroom can produce a report on a municipal zoning decision that directly affects fifteen thousand residents. If that article does not reach those residents because the algorithm favors high-engagement national content, the money paid has not fulfilled its social function.
Australian Anticipation
Australia had adopted comparable legislation in February 2021. The Australian News Media Bargaining Code had created similar pressure on platforms, with an arbitration mechanism in case negotiations failed. Meta reacted the same way: temporary blocking of news links, then return to the table after a text modification that reduced the group’s financial exposure.
The Australian result, three years later, is mixed. Large press groups concluded substantial agreements with Google and Meta. Some small independent publishers benefited from the mechanism unevenly, lacking the power to negotiate from a strong position. Australia subsequently strengthened the system to support small structures, regional zones, and, subject to implementation rules, diverse communities.
Canada is watching this trajectory. Canadian law includes specific provisions for minority-language media and Indigenous media, which number among the 425 beneficiaries in 2026. This is a notable design difference: acknowledging that not all local media are equivalent on the market, and that some fulfill social functions that the market alone will not finance.
Three Paths That the Canadian Assessment Opens
The CRTC report is an audit: it presents factual findings and signals areas of uncertainty. It makes visible an architecture of news financing that deserves explicit examination.
The first question concerns the absent platform. If Meta can exclude itself from the mechanism by removing news links, the law loses part of its reach. Discussions are underway in Canada on the possibility of broadening the obligation to platforms that benefit from journalistic content even without distributing it directly—for example, through artificial intelligence systems trained on press corpora. Collective bargaining on AI at work follows comparable logic: defining who captures the value of content produced by others, and how to redistribute a portion of it.
The second question concerns distribution. An annual fund of $100 million distributed among 425 media would represent an average of approximately $235,000 Canadian per beneficiary. But an average tells nothing about actual distribution, and the media that lost the most in the digital transition are not necessarily those receiving the most in the current mechanism. A revision of weighting criteria, toward greater weight given to the number of local journalists rather than audience, is defended by several community press associations.
The third path is more structural. Some media economics researchers advance the idea of coupling funding with distribution: making part of aid conditional on platforms committing to promoting local content. The idea is technically complex, notably because recommendation algorithms are commercial assets that platforms do not wish to subject to external regulation. But the logic is consistent with the stated objective: if the problem is invisibility, the remedy must also act on visibility.
Use of Funds by the 425 Newsrooms
The disintermediation strategy that some newsrooms have undertaken with these funds nonetheless has its own limit. Building a direct audience requires time, product development skills, and the capacity to retain readers in an environment where attention is already heavily solicited. These resources are precisely those lacking in the most fragile newsrooms. Media better equipped in personnel and digital expertise start with a structural advantage in this transition, which risks reproducing within the compensation mechanism itself the inequalities it seeks to correct.
The report documents payments and declared eligibility data, but not declared uses of funds by beneficiaries. The CCJ planned to collect additional data in subsequent years. For some small weeklies, the fund enabled preservation of one or more journalist positions.
That is concrete. A local journalist covering municipal councils, lower courts, and traffic accidents produces information that no one else will produce if that position disappears. The Canadian fund has preserved production capacity in areas where it was diminishing. It is a real result, even if partial.
Several beneficiary media also used part of the funds to develop direct newsletters and their own mobile applications, seeking to build audience relationships that do not pass through platforms. This is the direction several observers believe most promising by 2030: subscription and loyalty models that make local media less dependent on algorithmic distribution, whatever level of compensation they receive from platforms.
The tension between these two logics—taking money from platforms while trying to free yourself from them—is the thread the 2026 assessment leaves open. It will not be resolved by another law. It will be resolved, if it is resolved, by newsrooms capable of building a direct link sufficiently strong with their community to survive even when the algorithm ignores them.
