UAE R&D credits and Pillar Two: what changes in the spreadsheet, and what does not in the lab

1. Why this conversation is happening now

The UAE enacted its Corporate Tax framework in 2023 and, with Cabinet Decision No. 215 of 2025 and Ministerial Decision No. 24 of 2026, introduced a tiered R&D credit as Phase 1 of a broader innovation incentive strategy. The credit is non-refundable and sits against a 9% headline CT rate, which is itself below the 15% Pillar Two minimum rate threshold that large multinational groups are now subject to.

For most domestic UAE companies, the interplay with Pillar Two is straightforward: they are not in scope. For multinational groups with consolidated revenues above €750 million, the question becomes more nuanced. The treatment of the UAE R&D credit under GloBE rules, specifically, whether it reduces covered taxes and how it flows through the effective tax rate calculation, is material to any ETR model. Guidance on substance-based tax incentives under the OECD framework continues to evolve, and the treatment of non-refundable credits specifically is an area where adviser input is essential.

This article does not answer the ETR math; your Big Four team owns that. It answers an adjacent question that every CFO should put to the CTO: regardless of what the model says, do we have the engineering records to defend our entitlement to the credit, and our substance, if regulators ever ask?

2. A brief primer on the Pillar Two mechanics that matter

The OECD Pillar Two framework, sometimes called the Global Anti-Base Erosion rules, or GloBE, sets a 15% global minimum tax on the income of large multinational enterprises, applied jurisdiction by jurisdiction. Where a group's effective tax rate in a given jurisdiction falls below 15%, a top-up tax is charged, either by the jurisdiction itself (via a Qualified Domestic Minimum Top-up Tax, or QDMTT) or by the ultimate parent's jurisdiction under the Income Inclusion Rule (IIR) or Undertaxed Profits Rule (UTPR).

The UAE introduced its own DMTT, effective for fiscal years starting on or after 1 January 2025, which functions as a QDMTT and is designed to capture the top-up before foreign jurisdictions can. For in-scope MNEs operating in the UAE, this means their UAE tax burden will, at minimum, approximate the 15% GloBE floor when combining CT and DMTT.

Within this framework, tax credits matter because of how they are treated in the covered-tax calculation. Refundable credits generally do not reduce covered taxes (they are treated more like grants); non-refundable credits that reduce CT liability do reduce covered taxes and can therefore affect the ETR calculation. The UAE Phase 1 R&D credit, being non-refundable, falls into the latter category. Its interaction with the ETR and DMTT is precisely what finance teams are modelling.

3. What the modelling actually asks of your engineering team

Here is the practical consequence that often gets lost in the ETR conversation: the credit only exists if the underlying R&D expenditure is legitimate, documented, and pre-approved by the UAE R&D Council. A clean ETR model built on a weak engineering file is, from a risk perspective, no different from a house built on sand. The model assumes the credit; the audit tests whether you earned it.

This matters in both directions. If your advisers conclude that taking the UAE R&D credit in a given year creates unfavourable Pillar Two dynamics (for example, reducing covered taxes in a way that triggers top-up in another jurisdiction), your engineering team's documentation still has value. It establishes the substance that supports transfer pricing positions, protects against clawback risk if the credit is later challenged, and it positions the group well for Phase 2 of the UAE incentive programme, where refundability may be introduced and the Pillar Two dynamics would shift.

Conversely, if the model says the credit is worth claiming in full, a weak technical file creates exposure on multiple fronts simultaneously: the credit itself, the ETR position built on it, and any transfer-pricing arguments about where the R&D substance genuinely sits.

4. The substance question in Pillar Two is not separate from Frascati

There is a thread running through Pillar Two that many in-house teams have not yet connected to their UAE R&D preparations: substance. The GloBE framework includes Substance-Based Income Exclusions (SBIE) that carve out a portion of income attributable to genuine economic activity (payroll costs and tangible assets) from the income subject to top-up tax. While the SBIE calculation is not directly dependent on whether work qualifies as R&D under Frascati, the broader question of whether a UAE entity has real substance rather than a brass-plate operation is directly relevant.

Regulators and advisers will look at the same facts through two lenses. Through the UAE Corporate Tax and R&D credit lens: does this expenditure meet Frascati criteria, was it pre-approved, and are the people involved genuinely based in the UAE doing qualifying work? Through the Pillar Two lens: does this entity have sufficient economic substance, with real employees doing real work, to justify its position in the group's structure? Strong Frascati documentation is also strong substance documentation. The two are not the same analysis, but they draw on overlapping evidence.

5. What documentation looks like for an MNE hub in the UAE

Multinational groups with UAE-based R&D centres, typically in one of the free zones or on the mainland, face a documentation challenge that domestic-only companies do not. Their engineering teams often work alongside colleagues in other jurisdictions, on shared codebases, using shared infrastructure. The question of which costs are attributable to UAE-based R&D activity, performed by UAE-based staff, on projects approved by the UAE R&D Council, requires deliberate accounting that most project management tools do not provide by default.

A group that treats this as a year-end finance exercise, pulling together timesheets and cost allocations after the fact, will struggle. A group that builds UAE-specific project codes into its engineering workflows from day one, that links commits and tickets to pre-approved project identifiers, and that maintains contemporaneous records of who worked on what and when, will have both a defensible credit claim and a clean set of facts for any adviser modelling the Pillar Two position.

The evidence should include, at minimum: dated technical records linking personnel to approved R&D projects (commits, experiment logs, design documents), time allocation records distinguishing UAE-based R&D from other work, financial records tying costs to qualifying expenditure categories, and the R&D Council pre-approval correspondence for each project. For groups with transfer-pricing considerations, these records should be developed in coordination with the TP function from the outset, not retrofitted.

6. A hypothetical: US parent, UAE engineering subsidiary

Consider a fictional technology group headquartered in New York with a 200-person engineering subsidiary in Dubai Internet City. The group exceeds the €750 million revenue threshold and is therefore in scope for Pillar Two. The Dubai entity conducts genuine platform R&D (novel inference architecture, several multi-quarter projects with documented uncertainty and systematic experimental cycles). The group claims the UAE R&D credit on qualifying expenditure.

The finance team's Pillar Two model shows that the UAE R&D credit reduces covered taxes in the UAE entity, potentially affecting the ETR calculation. Depending on the magnitude of the credit, the structure of the DMTT, and the treatment agreed with advisers, the net ETR position after top-up may narrow the cash value of the credit relative to a standalone calculation. This is a real consideration and not one this article resolves.

What the article does note: the Dubai engineering team's records (architecture decision records, experiment logs, failure analyses) have value independent of how the ETR model resolves. They support the credit claim. They support the transfer-pricing argument that genuine R&D substance sits in Dubai. They support the SBIE calculation. And they protect the group in the event of any FTA or Council challenge. If the documentation is weak, all of those positions weaken simultaneously.

7. Phase 2 and refundability: why documentation built now pays off later

The UAE has structured Phase 1 as a foundation. Future phases of the R&D incentive programme are expected to introduce higher credit rates and, potentially, refundability for smaller companies or qualifying activities. If refundability is introduced, the Pillar Two dynamics shift materially: a qualified refundable tax credit is generally treated differently under GloBE rules than a non-refundable credit, and the ETR impact could diminish or disappear.

Groups that build robust Frascati-aligned documentation now are positioning themselves to claim fully across future phases without the scramble of recreating evidence that was never captured. The institutional knowledge of which projects qualified, which engineers worked on them, and what the experimental record looked like is genuinely valuable, and genuinely fragile if not maintained continuously.

8. What AutoDoc does (and does not do)

AutoDoc connects to the systems where engineering work already happens (Jira, GitHub, Linear, Notion, and others) and structures the evidence those systems generate into Frascati-aligned technical narratives. The output is a defensible file that your UAE tax advisers can review, that your R&D Council pre-approval application can draw on, and that you can use to answer an FTA examination without reconstructing history from memory.

We do not provide Pillar Two modelling, ETR analysis, DMTT calculations, or transfer-pricing opinions. Those belong with your international tax advisers, and any group in scope for Pillar Two should have specialist counsel across all of those areas before making credit claims.

What we offer is the engineering layer: contemporaneous, traceable, Frascati-structured records that make your advisers' work faster and your positions more defensible. The division of labour is the same reason strong SR&ED practices partner with us for Canadian claims. Documentation and tax strategy are related disciplines, but they are not the same job.