January 28, 2026 — The intellectual property of the world's largest multinational groups was, until very recently, located in places that bore little operational relationship to where the IP was created, used, or commercialised. Pharmaceutical patents originating in laboratories in Cambridge, Massachusetts, were licensed to entities in Bermuda. Software code written in Mountain View was held by entities in the Cayman Islands. Trademark rights for global consumer brands were resident in entities in Luxembourg or the Netherlands. The location of the intellectual property bore no necessary connection to the location of the engineers, scientists, or marketers whose work produced or commercialised it.

The architecture was constructed deliberately, over a period of decades, in response to the tax rules of the relevant jurisdictions. The architecture was legal, transparent in many cases to the relevant tax authorities, and produced effective tax rates on intangible income that were a fraction of the headline corporate tax rates of the jurisdictions where the underlying activities took place. The architecture was also, by the late 2010s, the principal target of the OECD's Base Erosion and Profit Shifting project. Pillar One and Pillar Two, the two pillars of the OECD's Two-Pillar Solution, were addressed substantially at this architecture.

By 2026, the picture has changed. This article describes where the intangibles of major multinational groups have actually ended up, what considerations drove the relocations, and what the intangibles transfer pricing landscape now looks like in practice.

Why intangibles became the central question

The location of intangibles was not always the central question of transfer pricing. For most of the post-1968 history of the OECD Transfer Pricing Guidelines, the central question was the pricing of cross-border tangible goods transactions, intercompany services, and intercompany financial flows. Intangibles were addressed but not at the centre of the framework.

What changed was the proportion of multinational profit attributable to intangibles. Through the 1990s and 2000s, the world's largest companies became progressively more intangible-intensive. Pharmaceutical companies derived an increasing share of profit from patents. Technology companies derived essentially all profit from software and platform code. Consumer goods companies derived growing share from brand value. The transfer pricing question that had once been about the price of widgets became, for the largest groups, a question about the location and pricing of intellectual property.

The location question and the pricing question are linked. If a multinational group locates its principal intangibles in a low-tax jurisdiction and licenses them to operating subsidiaries in higher-tax jurisdictions, the royalty rate on the licence determines the share of total group profit that flows to the low-tax holder. A higher royalty rate concentrates profit at the holder; a lower rate distributes profit to operating subsidiaries. The tax authorities of the operating subsidiary jurisdictions have, since the early 2000s, scrutinised these royalty rates aggressively. The largest international tax controversies of the 2010s — in pharmaceutical, technology, and consumer goods sectors — turned on whether the royalty rate paid by an operating subsidiary in country A to an intangibles holder in country B was arm's length, with country A's revenue authority arguing for a lower rate (more profit retained in country A) and country B's authority indifferent (since the holder paid little tax regardless).

The DEMPE framework and what it actually measures

The OECD's response to the intangibles question was the DEMPE framework, articulated in the 2017 revisions to the Transfer Pricing Guidelines. DEMPE stands for Development, Enhancement, Maintenance, Protection, and Exploitation. The framework provides that the entity entitled to the residual economic return on an intangible is the entity that performs the DEMPE functions, controls the DEMPE risks, and uses its assets to perform those functions — not necessarily the entity that holds legal title.

The framework was, in design, a substantial departure from prior practice. Pre-2017, the legal owner of an intangible was generally entitled to the residual return on it, with arm's length compensation paid to other group members for services rendered. Post-2017, the legal owner is entitled to the return only to the extent that it actually performs the DEMPE functions; if those functions are performed by other group members, the residual return follows the DEMPE performance regardless of legal title.

The implementation of DEMPE has been incremental and contested. The OECD published additional guidance in 2018 and 2020 addressing the application of the framework to specific intangibles — hard-to-value intangibles, marketing intangibles, and intangibles in restructurings. National tax authorities have absorbed the framework into their domestic transfer pricing practices at varying speeds. The German Federal Tax Court issued decisions in 2021 and 2023 that adopted DEMPE-influenced reasoning. The French administration has followed the framework in its 2024 examination procedures. The US Internal Revenue Service has had the framework on its agenda but has continued to develop its own intangibles doctrine through Tax Court litigation, with the Coca-Cola, Medtronic, and Eaton lines of cases providing the principal US doctrine.

The practical effect of DEMPE has been to make the location of intangibles less determinative than it was. Pre-DEMPE, the principal question was where the intangible was held. Post-DEMPE, the principal question is where the people are who perform the DEMPE functions. A multinational group that holds an intangible in Bermuda but performs all of the DEMPE functions in the United States has, under the DEMPE framework, allocated the residual economic return to the United States, regardless of the Bermuda legal title.

The migrations: actual cases through 2024 and 2025

The combined effect of DEMPE, Pillar Two, and the increasing aggression of national tax authorities has been to migrate intangibles to jurisdictions where the underlying functions are actually performed. The migrations have been more concentrated in the post-2022 period than in any prior comparable interval.

The most visible pattern has been the migration of US-developed intangibles from Bermuda, the Cayman Islands, and the British Virgin Islands to Ireland. The migration has been driven by a combination of factors: the OECD's pressure on the original holding structures, the introduction of the US Global Intangible Low-Taxed Income regime in 2017, and the substance requirements that Ireland has imposed for IP holding companies. By 2025, several major US technology and pharmaceutical groups had completed migrations of their principal IP assets from Caribbean holding entities to Irish holding entities, with the Irish entities staffed by genuinely substantive teams performing DEMPE functions.

The Irish migrations have not, however, ended in Ireland. Several groups that completed Irish migrations through 2018 to 2022 have, in the post-Pillar Two period, undertaken further reorganisations to rationalise the structure. The Irish twelve and a half percent corporate rate, combined with Pillar Two top-up to fifteen percent for in-scope groups, has reduced the marginal benefit of the Irish location. Some groups have consolidated multiple Irish entities into a smaller number of larger entities. Others have moved selected functions to Switzerland or the Netherlands where the operational ecosystem for specific activities is stronger.

The Swiss migrations have a different character. Switzerland's cantonal system produces effective tax rates that vary materially by canton, with a range from approximately twelve to twenty-one percent. The post-Pillar Two QDMTT brings in-scope groups to fifteen percent regardless of canton, but the operational and substance ecosystem of the cantons remains differentiated. Several pharmaceutical groups have concentrated R&D and IP holding in Basel and Zug; several technology groups have concentrated regional management in Zurich. The migrations have been substance-driven rather than rate-driven, and they reflect the new field of competition that Pillar Two has introduced.

The Singapore migrations are smaller in number but significant. Singapore's seventeen percent corporate rate, combined with the IP Development Incentive that provides reductions for qualifying activities, has positioned Singapore as the principal Asian alternative to Hong Kong for IP holding. Several groups with significant Asia-Pacific commercial operations have established or expanded Singapore IP holding entities, with the Singaporean entities staffed by genuinely substantive teams performing DEMPE functions for the Asia-Pacific region.

The Netherlands has retained significant IP holding activity, despite the curtailment of the Innovation Box regime. The Innovation Box, which provided a nine percent effective rate on qualifying intellectual property income, has been adjusted to operate within the Pillar Two framework. The post-2024 Innovation Box continues to operate but at an effective rate that, for in-scope groups, no longer drops below fifteen percent. The Netherlands' continued attraction is now operational rather than rate-driven: the country's substance ecosystem, treaty network, and legal infrastructure for IP holding remain strong.

The Pillar One overlay: Amount A and Amount B

Pillar One was the OECD's parallel reform addressed at the allocation of taxing rights, distinct from but conceptually linked to Pillar Two's minimum effective rate. Pillar One contains two components: Amount A, which would reallocate a portion of the residual profit of the largest hundred multinational groups to the jurisdictions where their customers are located, and Amount B, which provides a simplified return for routine distribution functions.

Amount A required a Multilateral Convention to enter force, signed by a sufficient number of jurisdictions to make the reallocation operational. The OECD published the Convention text in October 2023, but the instrument has not been formally opened for signature: deadlines have been extended multiple times through 2024 and 2025, and as of early 2026 the prospect of Amount A entering force in its current draft form is widely regarded as remote. The political opposition to Amount A in the United States Senate has been the principal obstacle, but other jurisdictions have also raised concerns about the design.

Amount B has had a different trajectory. The framework for Amount B was published in February 2024 and provides a simplified pricing methodology for routine distribution and marketing activities. The methodology is meant to reduce transfer pricing controversy in the lower-margin parts of multinational supply chains. Implementation by national tax authorities has been uneven. Several jurisdictions have adopted Amount B as a safe harbour from January 2025; others have indicated they will not adopt it. The status remains in flux at the date of this article.

The implication for intangibles is that Pillar One's reallocation of residual profit has not occurred. The pre-Pillar One transfer pricing framework, supplemented by DEMPE and overlaid by Pillar Two, remains the operative regime. The location of intangibles continues to determine, in significant part, where the residual profit on intangible-driven business is reported. The intensity of competition between jurisdictions for the location of intangibles is, if anything, higher in 2025 and 2026 than it was in the 2017 to 2020 period, because Pillar Two has converged effective rates without converging the location of profit.

Where the intangibles ended up

The 2025 picture of intangibles location is materially different from the 2015 picture. The principal changes are as follows.

Caribbean and Channel Islands holding entities have declined as a percentage of total intangibles location. The original Bermuda, Cayman, and BVI structures that dominated US technology and pharmaceutical IP holding through the 2010s have been substantially migrated. The residual structures are concentrated in groups whose IP is legacy rather than active: groups whose principal patents are mature, whose principal trademarks are stable, and whose DEMPE functions have largely terminated. Active IP development and active IP holding have moved elsewhere.

Ireland has absorbed a significant share of the migrating IP, but its share has stabilised rather than continuing to grow. The post-2024 Pillar Two environment, combined with the substance requirements that Ireland has implemented in connection with the Knowledge Development Box and other regimes, has made Ireland an attractive but no longer dominantly attractive location.

Switzerland, the Netherlands, and Singapore have absorbed share at Ireland's expense. Each offers a substance ecosystem suited to particular industries: Switzerland for pharmaceutical and chemical IP, the Netherlands for European regional management and consumer brand IP, Singapore for Asia-Pacific commercial operations. The differentiation among these three is increasingly operational rather than fiscal.

The United States has absorbed a smaller but significant share, particularly through the GILTI regime's interaction with the foreign-derived intangible income deduction. US-parented groups have, in selected cases, repatriated intangibles to the US to align legal title with DEMPE function performance. The repatriations have been driven less by tax optimisation than by audit risk management: a US-parented group whose principal DEMPE functions are conducted in California is, under DEMPE, substantially exposed to US transfer pricing adjustment regardless of where legal title sits, and the consolidation of legal title with DEMPE function performance reduces this exposure.

The United Arab Emirates has emerged as a small but growing destination for IP holding, particularly in connection with regional headquarters operations for the Middle East and Africa. The UAE's nine percent corporate tax, combined with the QDMTT for in-scope groups, places effective rates at fifteen percent for groups in Pillar Two scope and at nine percent for groups out of scope. The substance ecosystem in the UAE is in active development, with DIFC and ADGM providing the principal financial-services-oriented platforms and the federal Mainland providing operational infrastructure. The volume of intangibles holding in the UAE remains small relative to Ireland or Switzerland, but the growth rate is high.

The intangibles audit reality

The intangibles audit reality in 2026 is more aggressive and more technically sophisticated than at any prior point in the discipline's history. National tax authorities have built dedicated transfer pricing teams with intangibles expertise, supported by independent valuation specialists and access to the OECD's increasingly granular guidance.

The principal audit themes have stabilised. The first is the consistency between legal title and DEMPE function performance. The second is the reasonableness of royalty rates between holders and operating subsidiaries, tested against external benchmarks and against the residual return that DEMPE-performing entities can be shown to require. The third is the treatment of hard-to-value intangibles, where the OECD's specific guidance on the treatment of unanticipated outcomes has been absorbed into national practice.

An additional theme is the audit of intangibles restructurings themselves. Multinational groups that have migrated intangibles between jurisdictions have, almost without exception, faced subsequent audit of the migration itself: the valuation of the intangible at the date of transfer, the application of exit charges in the originating jurisdiction, and the treatment of any subsequent profit emerging in the destination jurisdiction. The audits have been resolved in favour of the administrations more often than not, with adjustments in the hundreds of millions or billions of dollars not uncommon.

The advisory function on intangibles transfer pricing has shifted accordingly. The principal advisory questions are no longer principally about the location of intangibles but about the documentation of DEMPE function performance, the defence of valuations, and the management of restructuring exposure. The advisory function is also increasingly forward-looking: the principal value of the function is in the design of structures that will withstand audit, rather than the resolution of structures designed in earlier eras.

The trajectory through 2030

The trajectory of intangibles transfer pricing through the second half of the 2020s is reasonably well-defined. Pillar Two will continue to converge effective rates. DEMPE will continue to determine where residual profit is allocated. The audit intensity will continue to grow. The principal locations of active intangibles holding will remain Ireland, Switzerland, the Netherlands, Singapore, the United States, and — increasingly — the United Arab Emirates. The dramatic migrations of the 2018 to 2024 period are largely complete; the next five years will be characterised more by refinement than by relocation.

The political question of whether Pillar One Amount A will eventually enter force will remain open. If it does, the picture changes materially: a portion of the residual profit of the largest hundred groups will be reallocated to customer jurisdictions, reducing the importance of intangibles location for those groups. If it does not, the current architecture stabilises in approximately its present form. The conditional path is wide.

What is no longer open is the question of whether the architecture of the 2010s will return. The Caribbean and Channel Islands holding entities, the Bermuda IP companies, the Cayman finance companies that defined a generation of multinational tax planning are not coming back. The next architecture, whatever it ultimately looks like, will be built on substance, on DEMPE, and on the floor that Pillar Two has established. The transfer pricing function inside multinational groups, and the advisory function that supports it, has adjusted to this reality. The adjustment is now substantially complete.