This is an attempt of an explainer article that captures what is going on globally around recent deal making activity to the best of my understanding.
The digital era has upended traditional notions of taxation in global trade. In the analog world, pre-trade world, corporate profits are typically taxed in the country where goods are manufactured. A German carmaker producing in Germany and exporting to China, for instance, would generally recognize most of its profits in Germany and pay corporate tax there, unless it had significant operations in China generating taxable profit.
However, this logic no longer holds true in the digital age. Digital services and intellectual property (IP) are taxed not where they’re consumed or where value is created, but where the IP is held – a location that can be easily shifted to low-tax jurisdictions. This disconnect between taxable profit location and economic activity stems from the intangible nature of digital goods and the flexibility surrounding IP ownership.
OECD accord
The OECD accord aimed to address this mismatch by ensuring large multinational companies in the digital sector pay taxes on profits based in countries where their customers are located, not just where the company claims its IP or headquarters. This represents a significant shift towards aligning taxation rights with the realities of a digitally interconnected global economy. A simple way to doing this is to apply a revenue principle.
An Irish Fudge?
Up to 2015, the infamous “Double Irish sandwich” tax avoidance scheme allowed multinational digital firms to channel vast revenues through Ireland, avoiding meaningful taxation in both the US and Europe. This loophole exploited differences in Irish and US tax laws, enabling companies to shift profits to low- or no-tax jurisdictions via Ireland. The scale of this avoidance was so significant that when Ireland closed the Double Irish arrangement in 2015, its GDP data for that year became distorted, reflecting the sudden onshoring of vast amounts of intellectual property and profits.
The implicit trade-off during this era was clear: societies accepted free digital infrastructure and the rapid expansion of digital ecosystems while allowing the marketization and monetization of user information. The EU’s General Data Protection Regulation (GDPR) emerged as a response, aiming to limit these practices and ensure “informed consent” in the digital economy.
OECD accord
As Ireland phased out the Double Irish and the OECD’s tax reform agenda advanced, the US – under Trump 1 in 2017 – slashed its corporate tax rate from 35% to 21% while introducing a low-rate minimum tax on foreign earnings. This was to make it cheap for big tech firms to repatriate their profits. In parallel, France and the UK imposed digital services taxes on revenue from tech giants, and several countries moved towards more granular, jurisdiction-level profit or revenue reporting, which could have created the data trace. In 2021, Biden signed on to the OECD’s 15% global minimum tax accord, but Congress blocked its implementation.
Prior to taking office, Trump had toyed with slashing the US corporate tax rate to 15%, matching the OECD floor. The implicit message was clear: Washington can undercut any country trying to tax American tech giants, leveraging its economic heft and security alliances as leverage. At the time, I urged policy makers around the world to unite to establish crude DST taxes on revenues as an alternative. This did not happen.
The EU/US “trade deal”
Fast-forward to the present, months into Trump’s second term, and it appears that the OECD’s hard-fought accord on taxing the digital economy is quietly being hollowed out or repurposed for US power projection. Signals from the G7 and EU suggest a willingness to move towards an alternative, such as a revenue-based minimum tax proposed by Scott Bessent. The EU similarly is exploring a broader large company tax. I am quite certain these are connected and may well have been part of the discussions between the EU and the US in Canada.
A tax on revenues makes sense, particularly for high-margin digital platforms whose marginal cost per user approaches zero and whose value scales through network effects and economies of scope. A revenue base is the correct concept for a tax base here. As AI continues to proliferate, the value may not be created in the development of more foundation models – but rather in the deployment of LLMs across society and economy broadly.
The money will be made in the deployment of AI in society, up until labor markets collapse. As long as inference can be localized (the US is pushing against this in its trade negotiations with an articulated aversion against data localization mandates), the tax base and the economic substance eventually can and will shift into the markets where revenues are generated.
The European Malaise
The lack of an integrated single market is a huge problem at present. Nevertheless, slowly but steadily pillars are coming together. There will be more and more support to build a healthy venture capital system and an accommodating tax regime, despite growing pains, we will see EU AI clusters emerge. And lastly, the EU is well positioned to gain. Why? The mere fact that AI will change how we interact with the internet, raises questions about trust, veracity of information, and societal stewardship. These are areas where the EU has traditionally been a leader.
If taxation ultimately boils down to the question of how much time people spend in different places by virtue of being owners of IP, measuring and arbitration are necessary to allocate taxation rights or assign tax residence.
The US is well-positioned to build this rail through its tech companies, thanks to control over cloud infrastructure, identity systems, payments architecture, app distribution – all remaining under American control. This gives Washington enduring leverage over economic dependencies, information flows, and ultimately, the political life of other societies. Even in a world with more competitors, the choke points of the digital age may still run through the United States.
So, it’s no wonder I fear Europe effectively capitulated on this issue. I fear that the “strategy” of playing the long game may not “work” in this setting, as any European tech stack will be solidly built by American tech companies or their (former) staff using US tech company rails.