How to Tax Digital Multinationals (Without Breaking the System)
The traditional tax system, based on the taxation of multinationals where value is produced, has gone into crisis with the advent of the digital economy. For over a decade, international organizations such as the OECD and the G20 have joined in the attempt to close the tax evasion loopholes linked to the legislative gaps in some systems and the absence of taxation in others, seeking to find an agreement on the allocation of the taxable income for multinational companies operating in the digital sector. Yet, despite ostensibly broad political consensus, a technical solution capable of satisfying all the various national interests has not yet been found.
A bargaining equilibrium point was the OECD-based global minimum tax, already adopted by the European Union and many other nations, with the aim of taxing all large multinational companies at an effective rate of at least 15%. In countries like Italy, however, compliance costs for companies sometimes exceed perceived tax benefits, and the effectiveness of the measure is curtailed by the absence of two key players, the United States and China. Another OECD proposal, aimed at reallocating part of the profits of multinationals to the countries where users and consumers reside, in order to reflect the added value they provide in the value chains of digital companies, is stalled. This multilateral vacuum has pushed many jurisdictions to adopt unilateral solutions, such as a digital services tax on revenues of technology companies. Such tax, however, is a cause of trade frictions with countries where Big Techs are based. Emblematic of this is the US federal bill, the so-called “One Big Beautiful Bill”, which in Section 899 calls for the black-listing of certain foreign DSTs and surtaxes of up to 20% on their income, flanked by the reactivated arsenal of tariffs as per ex-Section 301. Fiscal diplomacy thus becomes a tool of industrial policy, aggravating transatlantic tensions.
Generative Artificial Intelligence (AI) adds further complexity. Fiscal doctrine and policy-makers propose to tax algorithmic value chains, allocating the taxable amount based on the origin of the data used for training machines and local queries: an informational nexus which would complement physical presence. The Italian tax authority is in the vanguard, anticipating a pay-for-access model that values data as economic equivalent within the scope of VAT legislation, thus recognizing the exchange of “data for free services” as a taxable exchange. At the same time, AI also works to serve controls. In fact, machine learning can be used to identify incorrect behavior by taxpayers, for example by cross-referencing data relating to turnover, payment flows and network metadata. However, all this clashes with the limits set to government intrusion in personal data set by the GDPR, the AI Act, and domestic regulations on the protection of fundamental rights, which create a tradeoff between tax effectiveness and taxpayer protection.
In this scenario of great uncertainty, companies react through a form of strategic compliance, which involves collaboration with tax authorities, focusing on preventive dialogue with them to reduce uncertainty and sanctions, and investments in tax technology, thus transforming "fiscal data" — previously only recognized in terms of pure compliance — into a potential competitive asset.
We are therefore observing a rapidly evolving regulatory framework. The challenge for the future is to design an innovative tax system that is consistent with principles of equity, neutrality and proportionality, and manages to impose levies on the new digital wealth without stifling its expansion, so as to support social cohesion at a time when the revision of multilateral trade and fiscal relations has become paramount.
