On June 24, 2026, the European Commission published the EU Tax Omnibus alongside a recast of the Directive on Administrative Cooperation. For the first time in roughly a decade, a major EU direct-tax initiative subtracts obligations rather than adding them: the Commission estimates the package will cut business costs by about €8 billion a year, €3.3 billion of it pure administrative cost. The largest single measure, intra-EU withholding tax free promises of worth €5.3 billion a year, does not arrive until 2037. This article sets out what the Omnibus does, what it is worth, and why, in the new era, the benefit goes to the groups that plan for it rather than wait for it.

This article was written by Rafi Mardroos and Mario van den Broek of RSM Netherlands' International Tax and Supply Chain Management teams.

The shift in direction

For more than a decade, EU corporate tax policy moved in one direction only. The OECD base-erosion programme, the two Anti-Tax Avoidance Directives (“ATAD”): ATAD I’s interest-limitation, controlled-foreign-company (“CFC”), exit-taxation and general-anti-abuse rules, then ATAD II’s anti-hybrid-mismatch regime, nine successive layers of the Directive on Administrative Cooperation from DAC1 to DAC9, public country-by-country reporting and the Pillar Two minimum tax each added another layer of obligation. Tax functions responded with more people, more systems and more documentation. The administrative base of an international group grew thicker every year. The logic of the system was additive, and it was mandatory: each rule applied whether a company did anything to invite it.

The Omnibus breaks that pattern. Published as a single Council directive (COM (2026) 560 final), it is a simplification package, not a repeal: the anti-avoidance architecture stays largely in place, but existing tax rules (e.g. earnings stripping interest deduction limitation rule and intra-EU withholding taxes) are softened. The Commission puts the saving at roughly €7.9 billion a year and presents the package as part of its competitiveness agenda.

The Omnibus is a draft directive that amends six instruments at once: the Interest and Royalties Directive (2003/49/EC), the Parent-Subsidiary Directive (2011/96/EU), the Tax Merger Directive (2009/133/EC), ATAD1 (2016/1164), the Dispute Resolution Mechanism Directive (2017/1852) and the FASTER Directive (2025/50). It rests on Article 115 TFEU, which means the Council must adopt it by unanimity after consulting the European Parliament, i.e. the same legal base, and the same veto-prone path, that has slowed every direct-tax file of the last decade. Direct taxation has no harmonisation base of its own in the Treaties; unlike VAT and the other indirect taxes, which travel under Article 113, it can be approximated only through the internal-market clause of Article 115, and that route gives the Parliament a consultative voice while leaving each Member State a veto. The text can still change, and unanimity is itself the largest single risk to the timeline.

The saving is concentrated in a few measures. The intra-EU withholding-tax elimination accounts for about €5.3 billion; the interest-limitation reforms for more than €500 million, of which the third-party loan carve-out alone is roughly €430 million; the removal of the overlap between the CFC rules and Pillar Two for about €160 million; and the parallel DAC recast (COM (2026) 308 final) for more than €1 billion. The measures fall into four key items: withholding tax; ATAD (the CFC regime, a new research and development (“R&D”) allowance and the earnings stripping interest deduction limitation); restructuring and dispute resolution; and reporting, which we will discuss through below.

Withholding tax: a self-assessed exemption, in force from 2037

This is a key relief of the package. Today the Interest and Royalties Directive and the Parent-Subsidiary Directive exempt qualifying intra-EU payments only above a participation threshold: a 25% holding for interest and royalties, a 10% holding for dividends. Moreover, Member States operate upfront relief procedures and clearances that a payer must complete before it can be generally assured that it will obtain a tax benefit (e.g. withholding tax exemption). Each exemption also presupposes that the recipient is the beneficial owner of the income, and a Member State may still require the holding to have been maintained for up for a certain period.

The draft Omnibus aims to soften these rules. It removes the minimum participation requirement from both directives so that withholding tax would be exempt regardless of percentage of shareholding. EU member states are also no longer allowed to demand certificates/permits in relation to the withholding tax exemption. Instead comes self-assessment where the taxpayer determines entitlement itself, subject to control by the tax authority after the event and to the existing anti-abuse rules. 

Moreover, where entitlement cannot be established at source (typically for listed securities held through nominee accounts) the relief-at-source and quick-refund procedures of the FASTER Directive apply. In addition, an additional withholding tax exemption is also granted to pension institutions.

Noteworthy is also that where interest or royalty payments are made to a recipient in a non-EU member state that would qualify as a low-taxed / non-cooperative jurisdiction lied at source, the source Member State in principle must either levy withholding tax or deny the deduction (unless pillar 2 taxation applies).

The constraint is timing. The Commission attributes about €5.3 billion of the annual saving to this measure, but most of the Omnibus applies from 1 January 2029 while the intra-EU withholding-tax softening is deferred to 1 January 2037; the present thresholds and procedures continue until then. The reversal of burden, however, from the day the exemption is self-assessed under the new 2037 regime, the taxpayer carries the burden of showing entitlement if the position is later opened by the relevant tax authority.

ATAD: a narrower CFC regime and a new R&D allowance

ATAD is narrowed at one end and extended at the other. On the CFC rules, the package removes a source of friction: groups within the scope of Pillar Two can be excluded from CFC taxation ending an overlap that would potentially create double taxation (CFC levy and Pillar 2 levy). Small and medium-sized groups, as defined in the Accounting Directive (2013/34), may be carved out entirely from the CFC rules. For the groups that remain, the optional Model B is deleted, leaving Model A as the single EU approach. 

At the other end, ATAD1 introduces, for the first time, an incentive: a minimum standard for the tax treatment of investment in R&D-related tangible assets. As such, ATAD1 is being repurposed to a directive that does more than only counteract tax avoidance. The mechanism for the R&D incentive is immediate expensing of qualifying capital expenditure on designated assets (e.g. plant, machinery and other tangible assets) used directly for R&D, deductible either in the year it is incurred or across any of the four following tax periods. The assets must be used for R&D for at least three years, and balancing charges apply on early disposal. 

The EU has also considered that the above-mentioned R&D allowance should depress the taxpayer’s fiscal EBITDA for earnings stripping purposes (lower fiscal EBITDA could result into more interest deduction limitation). As such, any R&D allowance expensing as described above may be disregarded for earnings stripping purposes which would be favourable for the taxpayer.

Interest limitation: a more investment-friendly Article 4

Article 4 of ATAD1 receives the most detailed rework in the package, with the aim to better facilitate financing within the EU. The 30% of EBITDA ceiling becomes a fixed figure that Member States can no longer reduce, ending a fragmentation in which national caps had drifted below the directive floor. Low-risk third-party loans, i.e. financing from unrelated lenders used to fund the borrower’s own activities rather than on-lent within the group fall outside the rule entirely; the Commission treats this as the single largest interest-limitation saving, about €430 million a year. The €3 million de minimis allowance becomes mandatory and is indexed to inflation from 1 January 2032. 
 

Restructuring, disputes and reporting

Four further changes matter for groups that reorganise, litigate or report. 

• The Tax Merger Directive is realigned with EU corporate law derived from the Mobility Directive (2019/2121). This means that the Tax Merger Directive’s roll-over relief would also apply to other forms or reorganization such as cross-border conversions.
• The General Anti-Abuse Rule in Article 6 of ATAD is confirmed to apply to all taxes, including withholding taxes and Pillar Two top-up taxes.
• The imported hybrid-mismatch rules are deleted as disproportionately complex for the protection they delivered. The imported-mismatch rule (Article 9(3) of ATAD II) was the most demanding of the anti-hybrid provisions, obliging a taxpayer to trace a deduction or non-inclusion arising entirely outside the Union and neutralise it through an otherwise ordinary intra-group payment, 
• The EU Dispute Resolution Mechanism (“DRM”) is an EU Directive that aims to achieve efficient resolution in relation to disputes related to the interpretation and application of tax treaties. The DRM will be adjusted to address procedural shortcomings that have often delayed or prevented a speedy dispute resolution. The aim is to help business resolve their tax disputes more efficiently in line with the original purpose of the DRM.

Operating in the new era: a transition map

Running alongside the Omnibus, the recast of the Directive on Administrative Cooperation consolidates DAC1 to DAC9 into a single instrument and trims the reporting estate. It removes the DAC6 cross-border-arrangement reporting obligation for the roughly 3,000 multinational groups already inside Pillar Two (about €300 million a year on its own) replaces duplicative country-by-country and top-up-tax notifications with a single filing, worth a further €260 million, and raises the DAC7 reporting threshold to €3,000. The reporting reversal is the same as the rest: less is asked, but what is asked must be right the first time, because the consolidation strips out some of the redundancy that used to catch errors. The dates govern the planning, and they are staggered.

The burden reversal changes how a position is defended. With prior authorisation gone, a group that takes the withholding exemption must be able to evidence entitlement under later examination, so contemporaneous documentation and a clean anti-abuse analysis carry weight they did not carry when a Member State signed off in advance. The default reversal rewards the same discipline from the other side: the R&D allowance, the third-party-loan treatment and the time-limited defence exclusion each require a decision about whether and how to claim, several carry deadlines, and modelling the position before capital is committed will usually return more than the routine compliance the package removes. Because the benefits are elective and the clock is staggered across 2029, 2032 and 2037, the practical task is not a single project but a sequence of windows, each with its own posture. The map below sets out what changes in each window, and what a group should be doing now to be ready when it does.
 

So, is there a new era in tax?

Two practical conclusions run through the transition map. The first concerns timing. The withholding-tax thresholds in the Interest and Royalties Directive and the Parent-Subsidiary Directive, and the upfront relief and clearance procedures that accompany them, remain in force until 1 January 2037. A group that dismantles those procedures in anticipation of the new self-assessment regime forfeits relief in the intervening years, and on a financing structure of any size that is the most expensive error the package makes available. The second concerns the character of the change. The Omnibus does not remove work so much as relocate it: pre-clearances become evidence files; mandatory rules become elective reliefs that a group must decide whether and when to claim; and the annual compliance cycle gives way to a staggered set of deadlines in 2029, 2032 and 2037. The groups that extract the most from the package will be those that treat it as a portfolio of positions to be claimed, evidenced and timed, rather than a simplification to be noted once and filed.

None of this arrives as relief in the near term. A simplification package still must be read against a group’s own flows, mapped onto its structures and modelled before it saves a euro, and in the first years it adds to the workload rather than reducing it. The pre-2037 withholding regime continues to run while the new logic is learned alongside it; elections such as the immediate expensing of R&D assets under the amended ATAD1, or the third-party-debt carve-out in the recast Article 4, must be weighed against capital expenditure not yet committed; and processes designed to obtain clearance have to be rebuilt to produce evidence. The legislative history counsels caution on the timetable. Direct-tax files proceed by unanimity under Article 115 TFEU, seldom survive the Council unaltered, and tend to arrive later and in amended form; the Commission’s own estimate places the largest single component of the saving, the €5.3 billion attributed to intra-EU withholding-tax relief, no earlier than 2037. The direction of travel is welcome. But a group that treats the Omnibus as relief it can simply wait for will find that, in the interim, it has acquired one further body of rules to understand and apply.

That pressure falls on functions with little capacity to absorb it. For a decade, each new obligation was met in principle by adding people; in practice, headcount has rarely matched the pace of rule-making, and the same teams now carry a heavier statutory base, a public country-by-country and Pillar Two reporting estate, and minimum-tax computations that did not exist five years ago. To ask those teams also to identify and claim elective reliefs, to assemble evidence files capable of surviving examination, and to model investment decisions before capital is committed, without revisiting how the function is resourced, is to expect the benefits of the new regime while withholding the means to secure them.

The most consequential development of this period may therefore lie outside the directive altogether. For the first time, a tax function can meet higher and qualitatively different demand without a corresponding increase in payroll, because the necessary capacity can now be procured by the task rather than by the post. Applied with discipline, artificial intelligence can read a directive against a group’s own intra-EU flows, prepare a first draft of an evidence file, surface the elective reliefs that a fully occupied team would otherwise miss, and keep the transition calendar under continuous review. The implication for an international tax function is not that it should add bodies, but that it should expect more of its systems and its advisers: technology-augmented capability in place of additional hours, with experienced professional judgement reserved for the questions that warrant it.

A second force points in the same direction. Tax is increasingly something on which a company is assessed, not merely something it remits. Transparency and governance expectations, public reporting, and carbon-pricing mechanisms such as the Carbon Border Adjustment Mechanism (CBAM) have drawn environmental and reputational considerations into the tax base itself, and the data a function must now command extends well beyond the corporate return. A team already operating at the limit of its capacity is being asked to work across more dimensions at once, which only strengthens the case for an operating model designed around leverage rather than the sheer volume of hours available.

So, is there a new era? There is, though not the one the headline saving of roughly €8 billion suggests. The Omnibus is not itself the new era; it is a test of how well a function is prepared for it. The reductions it offers are real, but they are deferred and elective, and they will accrue to the groups that have already organised themselves around intelligence, evidence and continuous adaptation, the same groups that will absorb the next obligation, in whichever direction it runs, without disruption. The durable shift is from a function that keeps pace with its obligations to one that extracts value from regulatory change of any kind; a function built on that basis is largely indifferent to whether the next file adds requirements or removes them.

This is the work RSM undertakes with international tax functions: not only reading the Omnibus and placing each measure on the transition calendar but building the operating model the new environment assumes. In practice that means valuing the elective reliefs against a group’s actual EU investment, preparing the evidence files that self-assessment now requires, and integrating the ESG and reporting data on which the modern tax base draws, supported by an AI-augmented method developed within our own practice that allows a stretched function to do more without simply adding to headcount.

Groups should also weigh the second-order consequences of the package. Its stated purpose is to strengthen EU competitiveness, and its instruments are designed to change behaviour, not merely to reduce cost. The R&D allowance may encourage research activity, and the tangible investment that accompanies it, to return to the Union; the resulting cross-border reorganisations will need to be structured efficiently for both tax and operational purposes. The more lenient earnings-stripping regime is intended to ease the financing of the economy; where a group responds by taking on additional third-party debt, the analysis cannot end at the interest-limitation position but must extend to the wider tax consequences of that financing.

Download the full report

RSM is a thought leader in the field of international tax consulting. We provide frequent insights through training and the sharing of thought leadership, based on our detailed knowledge of industry developments and practical applications gained from working with our clients. To map your intra-EU flows against the 2037 withholding regime, place your group in the new CFC bands, value the R&D and interest-limitation reliefs against planned EU investment, and build the evidence files the self-assessment era requires, please contact one of our consultants.

This article reflects the proposals published by the European Commission on 24 June 2026 (COM (2026) 560 final and COM (2026) 308 final). They remain subject to the special legislative procedure under Article 115 TFEU, including unanimous adoption by the Council, and may change before adoption and national transposition. The simulations are illustrative, use rounded assumptions and are not a substitute for advice on a specific group. This is general information and not specific tax advice.