A small country's bid to tax the rentiers of the internet

IN 2021 Australia became the first country to attempt to force big tech to pay for news. It was an odd experiment, born of the conviction that Google and Meta were extracting value from journalists without paying for it. The experiment initially worked. Google and Meta together paid roughly 250m Australian dollars ($155m) a year to local publishers. Then Meta found a loophole: it stopped renewing its deals in early 2024 and simply removed and de-prioritised news from its platforms. The code's threat - forced arbitration - could not reach a company that refused to carry news at all.

Two and a half years later, the Australian government has unveiled a successor that is designed to close that loophole. The News Bargaining Incentive, drafted in April 2026 and still in the consultation phase, would impose a 2.25% levy on the Australian revenue of digital platforms that fail to strike commercial deals with news publishers. The levy applies to social-media platforms and search engines with more than 250m Australian dollars in local revenue and, respectively, five million or 10 million users. By that test, Meta, Google and TikTok would be caught. Microsoft, Snapchat and OpenAI would not.

The clever bit is the offset. Platforms that pay publishers directly can deduct 150-170% of the deal value from their levy liability. Every dollar paid to a newsroom saves more than a dollar in taxes. If a company makes deals, its effective charge drops to roughly 1.5% - cheaper than the 2.25% flat rate. The design is a carrot-and-stick that has been carefully weighted so the carrot is always larger. Any revenue collected under the levy would be passed to news organisations, weighted by the number of journalists they employ, with extra favour to regional, small and multicultural outlets. The government would keep nothing.

On paper the mechanism is elegant. In practice it resembles a digital services tax. The distinction matters because the political and legal risks of such taxes are no longer theoretical. The Trump administration has already labelled the proposal "foreign extortion". US lobby groups, including the Computer and Communications Industry Association, have called for "targeted trade remedies" in retaliation. Meta has argued the scheme violates Australia's free-trade agreement with the United States, and the National Foreign Trade Council, which represents American multinationals, has agreed.

To be sure, these trade-law objections deserve scrutiny. The scheme singles out a handful of predominantly American firms while exempting others offering comparable services. That is the sort of discrimination that free-trade national-treatment clauses are designed to prevent. And Meta has a point about a different kind of fairness: most publishers share their content on social platforms voluntarily, because the audience reach and indirect advertising benefits are real. A government-mandated transfer of wealth, as Meta puts it, does not look the same as a market price.

Yet the deeper problem is not whether the levy is perfectly calibrated. It is that the platforms' rents are real, the decline of local journalism is real, and the original code was structurally toothless. Australia has lost around 166 news outlets since the bargaining code was first introduced in 2021, according to the Public Interest Journalism Initiative. Regional and remote areas have accounted for 60% of those closures, despite only 28% of Australians living outside metropolitan centres. Meta's $70m-a-year exit alone represented a sudden withdrawal of funding that smaller publishers had come to rely on. The incentive model is an acknowledgement that voluntary cooperation, backed by the threat of arbitration, does not work when the party with the most power decides that walking away is cheaper than paying.

The reason the original code collapsed is not hard to see. It required platforms to carry news in order to be subject to its rules. Meta removed news and the code's leverage evaporated. The new levy applies regardless of whether a platform hosts news content. That is the structural improvement: the obligation is no longer conditional on the platform's willingness to be a distributor.

But the design choices raise questions of their own. Excluding LinkedIn and Apple because they employ editorial teams internally is a tidy distinction, but it means the scheme's coverage depends on a somewhat arbitrary line between platform and publisher. Excluding AI firms entirely is politically easier but arguably more consequential. As generative models scrape news content for training data, the same extraction problem that drove the bargaining code may reappear in a new guise. The government says it will manage that issue separately, which is honest but also suggests the first shot may prove the opening skirmish.

The most significant constraint is geopolitical. All three of the platforms captured by the levy are American. The United States has already withdrawn its support for the OECD's global tax framework, and the Trump administration has repeatedly targeted digital services taxes as discriminatory. If Australia proceeds, the risk is not that the United States will retaliate immediately - it is that it will add the NBI to a growing list of grievances and use trade leverage elsewhere. Prime Minister Anthony Albanese has not flinched, insisting the decision is based on Australian national interest. But sovereignty is not infinite when trading partners control infrastructure.

The legislation is still delayed. The Albanese government has not yet introduced it to parliament after the May consultation closed, despite promises to table it before the July winter recess. Media chiefs, including Michael Miller of News Corp Australasia and Rebecca Costello of Guardian Australia, have sounded the alarm that further delays risk more belt-tightening and closures. The hesitation is understandable: the government is weighing domestic urgency against the risk of an international trade clash. That calculation is not a weakness; it is the nature of the problem.

The better outcome is for platforms to enter deals directly, as the NBI is designed to encourage. The offset structure makes that the rational choice. Google, which already has agreements with more than 200 outlets, may find the transition less jarring. Meta, which has spent the past two years arguing that the entire premise is wrong, is less likely to cooperate willingly. TikTok is a wild card: it has not yet responded to requests for comment, and its Australian footprint is large enough to matter.

What the case reveals is less about journalism and more about the distribution of power in the digital economy. Small countries are discovering that their users are the raw material from which global platforms extract advertising revenue. The platforms argue they provide free distribution and that publishers benefit. That is not entirely false. But the bargaining code showed that when the implicit bargain - free audience reach in return for free content - becomes unsustainable for the content producer, the party with fewer alternatives loses.

Australia's approach is not flawless. The selective coverage invites trade challenges. The exclusion of AI firms is an open wound. And there is no guarantee that 200-250m dollars a year will arrest the contraction of regional and remote journalism, which is driven by decades of advertising collapse as much as by platform extraction.

The aim should be to preserve a news ecosystem that is not dependent on government administration or platform benevolence. The NBI, if it survives the trade dispute it has already provoked, moves in that direction. The first task is to pass it. The harder one is to make sure the next round - AI, search diversification, the emergence of new platforms - does not leave the same problem unsolved. That bargain is not finished.