UAE R&D tax credit 2026: the guide for CFOs, heads of tax, and CTOs

1. Legal basis and effective date

The UAE R&D tax credit is established by Cabinet Decision No. 215 of 2025 and operationalised by Ministerial Decision No. 24 of 2026, implementing a credit mechanism within the UAE Corporate Tax regime. The credit builds on amendments introduced by Federal Decree-Law No. 28 of 2025, which updated the Corporate Tax Law to accommodate qualifying incentive schemes.

The regime applies to tax periods commencing on or after 1 January 2026. For most UAE-resident entities on a calendar tax year, this means the first claimable period runs through December 2026, with filings following in 2027.

One observation that practitioners from other jurisdictions consistently make: when Canada introduced SR&ED reforms or the US refined its §41 regulations, the headline rate was rarely the most consequential change. What determined outcomes was whether companies had the project governance and documentation discipline in place when the rules became enforceable. Year one of a new credit programme is the best time to build those habits, not year three when the first challenge letters arrive.

2. Who can claim, and who cannot

The credit is available to UAE-resident juridical persons and foreign enterprises with a UAE permanent establishment that perform qualifying R&D activities, provided they fall within Corporate Tax or Top-up Tax scope. Entities that have elected Small Business Relief, or that are otherwise not subject to Corporate Tax or the Domestic Minimum Top-up Tax, are generally excluded from the regime based on publicly available summaries.

Qualifying Free Zone Persons present an additional layer of complexity. The interaction between QFZP status, which ordinarily confers a 0% rate on qualifying income, and the R&D credit mechanism requires careful analysis. Whether a QFZP must be subject to 9% Corporate Tax on the relevant income, or subject to Top-up Tax, in the period of spend is a specialist modelling question that your tax adviser needs to address specifically for your entity structure.

3. The Frascati activity test in plain English

The UAE credit does not define qualifying R&D in its own vocabulary. Instead, it adopts the framework of the OECD Frascati Manual, which is also the eligibility standard behind Canada's SR&ED programme, the UK's R&D relief regime, and most other national innovation incentives. Under Frascati, qualifying work must be novel, creative, uncertain at the outset, pursued through a systematic method, and productive of knowledge that is in principle transferable or reproducible.

In practice, the novelty criterion asks whether your project advances beyond what was already publicly known in the relevant technical field. A team implementing a well-documented architecture is not doing novel R&D, even if the implementation is complex. A team designing a new approach to a problem where no established solution exists in the literature is doing novel R&D, provided the other criteria are also met.

The uncertainty criterion is the one that most often determines outcomes in review. Technological uncertainty means that, at the time the project began, it was genuinely unknown whether the objective was achievable, how it might be achieved, or how long it would take, and that this uncertainty could not be resolved by a competent practitioner applying known methods. The test is not about project risk or market uncertainty. It is about a specific epistemic gap: the answer was not available in the state of knowledge at the time. Documents that capture that gap at project initiation, not reconstructed at year-end, are the foundation of every defensible claim.

Activities in social sciences, humanities, and arts are excluded. Routine quality control, standard testing, and work whose outcome is predictable from known methods are also excluded. The boundary between qualifying investigation and skilled professional work is where most disputes arise, and where contemporaneous documentation makes the decisive difference.

4. Tiered credit mechanics with a worked example

The credit operates progressively across bands of qualifying expenditure, with each band carrying a different rate and a minimum average headcount of UAE-based R&D staff. You must satisfy both the spend level and the headcount requirement for each band; if your headcount falls short of a higher band's threshold, that band's rate does not apply regardless of your spend.

Based on publicly available summaries of Ministerial Decision No. 24 of 2026, the illustrative structure is as follows. The first AED 1,000,000 of qualifying spend attracts a 15% credit, with a minimum of two qualifying R&D staff. The band from AED 1,000,001 to AED 2,000,000 attracts 35%, with a minimum of six staff. The band from AED 2,000,001 to approximately AED 5,000,000 attracts 50%, with a minimum of fourteen staff. The maximum credit across all bands is commonly cited at around AED 2,000,000 per entity or tax group per period. Always verify exact figures against the official Arabic and English decision text.

A worked example

Consider a fictional Dubai-based technology company, NexGulf Analytics, with AED 2,400,000 of qualifying R&D expenditure in the period and an average of eight UAE-based R&D staff.

The first band applies straightforwardly: AED 1,000,000 at 15% produces AED 150,000 of credit. Eight staff exceeds the two-staff minimum comfortably. The second band also applies: AED 1,000,000 at 35% produces AED 350,000. Eight staff exceeds the six-staff minimum for this band. The remaining AED 400,000 falls into the third band, which would attract 50%, but that rate requires fourteen staff, and NexGulf has only eight. Under publicly described rules, the spend in the third band steps down to the highest rate for which the headcount is satisfied, which in this case would be 35%. The remaining AED 400,000 at 35% produces approximately AED 140,000.

Total illustrative credit: approximately AED 640,000. The important planning insight from this example is that headcount decisions have a direct impact on the marginal credit rate. Adding six qualifying R&D staff would push NexGulf into the 50% band for the third slice, turning AED 140,000 into AED 200,000. For a company modelling its UAE engineering headplan, the R&D credit structure is a material input to that decision.

5. R&D Council pre-approval and the Tawwer portal

The element of the UAE programme that most surprises practitioners from other jurisdictions is the pre-approval requirement. In Canada and the United States, R&D credits are self-assessed and claimed retrospectively through a tax filing, with the regulator reviewing after the fact. The UAE programme requires companies to obtain approval from the UAE R&D Council for each qualifying project before qualifying expenditure is incurred, or at the latest concurrent with work commencing.

Applications are submitted through the Tawwer portal, the UAE government's digital platform for this programme. Approvals are generally project-specific and valid for one tax year, meaning multi-year R&D initiatives require renewal applications for each subsequent period. For a company with a rolling portfolio of R&D projects, this creates a governance requirement that should be embedded into quarterly planning, not treated as an annual administrative afterthought.

A strong pre-approval application describes the project's technical objectives clearly, articulates what is genuinely unknown at the outset and why, explains the planned investigation methodology, and specifies the UAE-based personnel who will conduct the qualifying work. The technical section should read like a Frascati project brief, not a product roadmap. Advisers familiar with the programme consistently note that applications that fail to engage with the eligibility criteria in the technical narrative are more likely to face delays or rejections.

Our dedicated guide to the R&D Council pre-approval process covers the application structure, renewal mechanics, and how to handle mid-year project changes.

6. Qualifying costs, caps, and double-dipping rules

The qualifying cost categories commonly described in professional firm analyses include UAE-based R&D staff costs under the entity's direction, with a 30% overhead uplift on staff costs in many summaries. Consumables used directly in qualifying R&D activities are generally included. UAE subcontracting costs are potentially eligible, subject to arm's-length pricing and anti-chain restrictions that prevent the same expenditure being claimed at multiple levels of a group structure.

Cost contribution arrangements and certain capitalised R&D-related costs may also fall within the qualifying scope, depending on the specific facts and how the official guidance is applied. Government-funded R&D is generally excluded, as is expenditure that attracts relief under other incentive schemes. The double-dipping prohibition is consistent with international practice and applies both within the UAE and across jurisdictions.

A minimum qualifying expenditure threshold per project per tax period is commonly cited in professional commentary at approximately AED 500,000. Verify this figure against the official legal text before relying on it for project scoping decisions.

7. Documentation, retention, and filing

The credit is claimed through the UAE Corporate Tax return, or the Top-up Tax return where applicable. The supporting documentation required at filing typically includes evidence of R&D Council pre-approval for each claimed project, senior management declarations confirming the accuracy of the claim, detailed expenditure breakdowns attributing costs to qualifying activities and personnel, and audited financial statements where required by the relevant regulations.

Both technical and financial records must be retained for seven years from the end of the relevant tax period. For an engineering team, seven-year retention means maintaining immutable archives of the records that form the technical file: ticket history with timestamps, commit records, experiment logs, design documents, load-test results, and postmortems. If your current data retention policies allow engineers to delete repositories or project management data on a shorter cycle, those policies need to be updated before you make a credit claim that depends on those records surviving a review.

The financial records are equally important. Payroll records linking specific employees to specific R&D projects and activities, timesheet records showing time allocation, cost allocation memos, subcontractor invoices, and consumables receipts all need to be structured and retained in a form that supports the credit claim rather than merely existing somewhere in the accounting system.

8. How UAE compares to SR&ED and US §41

The Frascati eligibility framework creates a common conceptual foundation across the UAE credit, Canadian SR&ED, and US §41. In all three, the fundamental question is whether the engineering work involved genuine technical uncertainty, a systematic approach to resolving that uncertainty, and an advance in knowledge that was not available from the existing state of the art. The vocabulary differs across the three regimes, but the underlying engineering behaviour being documented is the same.

The most significant structural difference is the UAE's pre-approval requirement. Both SR&ED and §41 are entirely self-assessed, with government review occurring after filing. The UAE gate at the project level, before spending begins, changes the compliance cadence fundamentally. It also, in practice, tends to improve the quality of the technical file, because the eligibility analysis that other jurisdictions defer to year-end must be done upfront.

The credit mechanics differ in ways that matter for modelling. SR&ED offers an investment tax credit that is refundable for Canadian-controlled private corporations below the expenditure limit, making it a direct cash benefit for smaller companies. US §41 operates on an incremental base against a historical comparison period, creating a different calculation structure. The UAE's tiered-rate, headcount-gated approach is distinct from both, and requires a different kind of headcount and spend planning analysis.

9. Two detailed hypotheticals

Hypothetical A: enterprise SaaS, Dubai (fictional)

A fictional 40-person engineering team at a Dubai platform company is working to reduce p99 API latency from 120ms to under 50ms under Gulf-region peak load conditions. The team identifies this as a novel problem because the specific combination of their data characteristics, hardware constraints, and latency requirements rules out the standard caching and batching techniques documented in the literature.

Over six sprints, three distinct architectural approaches are tested and fail to meet the target under realistic load. A fourth approach, combining a novel prefetching strategy with a different queue topology, eventually achieves the target. The technical file for this project is not the slide deck summarising the outcome. It is the six sprints of load-test results, each documenting what was measured, why the approach failed, and what that failure implied about the design space. It is the architecture decision records written at the start of each sprint, stating the hypothesis being tested and the criteria that would constitute success or failure. It is the postmortem for each failed approach, specific about what the measurement showed.

A reviewer applying Frascati criteria would find novelty in the problem framing, uncertainty in the documented failure record, systematic investigation in the structured sprint methodology, and transferability in the postmortems. The same file would satisfy a CRA technical reviewer applying SR&ED criteria, with minimal reformatting. That is the value of building documentation from the engineering reality, not constructing a narrative around it.

Hypothetical B: industrial R&D, Abu Dhabi (fictional)

A fictional materials science group in Abu Dhabi is developing a new aluminium alloy formulation for structural components intended to operate in extreme desert heat-cycling conditions. Existing alloy compositions documented in the literature do not meet the required fatigue-life specifications under the temperature cycling profiles characteristic of Gulf outdoor applications. The team runs a structured programme of 23 furnace trials across seven alloy variants, varying composition and heat treatment parameters systematically.

The evidence for this project is the trial matrix itself: batch records with lot numbers, parameter setpoints, operator notes, and deviations for each trial; metallographic analyses and fatigue-test results; SPC charts showing how properties evolved across the trial sequence; and the decision record explaining which variants were eliminated and why, based on the measurement data. The R&D ends when the team identifies an alloy composition that consistently meets the specification. The scale-up and production work that follows is not R&D under Frascati criteria.

This kind of manufacturing R&D tends to produce naturally strong evidence because the measurement records are contemporaneous by default. The challenge is not creating the evidence but structuring it, connecting each trial's data to the specific technical question being investigated, in a way that reads as systematic inquiry rather than quality control.

10. Where AutoDoc fits

AutoDoc connects to the engineering and project management tools where R&D work already happens (GitHub, Jira, Linear, Confluence, and others) and structures the records those systems generate into Frascati-aligned technical narratives. For manufacturing and industrial teams, we work with structured exports from MES and QC systems. The output in both cases is a project-level evidence file that your UAE tax advisers can review, that your R&D Council pre-approval application can draw on, and that an FTA examination can be answered from without reconstructing history.

We do not provide tax advice, eligibility determinations, or filing services. Those belong with qualified UAE advisers. Our role is the documentation layer: the engineering records, structured contemporaneously, that make every other part of the process faster and more defensible.

The same model drives our SR&ED work in Canada, where contemporaneous documentation consistently outperforms retrospective reconstruction in CRA technical reviews. You can read about how Conceptinero improved SR&ED efficiency with AutoDoc, and then consider how the same operational workflow applies to your UAE projects through the UAE R&D hub.