New Zealand R&D Tax Incentive for AI Companies 2026

Detailed Program Overview

The New Zealand R&D Tax Incentive represents one of the country's most significant investments in fostering innovation and technological advancement. Administered by Inland Revenue New Zealand (IRD), this program was designed to address the critical need for increased research and development activity across New Zealand's business sector, particularly in high-growth industries like artificial intelligence and technology.

The incentive operates on a fundamental principle: reducing the financial risk associated with genuine innovation work by providing companies with a substantial tax credit for eligible R&D expenditure. Unlike traditional grants that require pre-approval and competitive selection processes, the R&D Tax Incentive functions as an entitlement program—companies that meet the eligibility criteria and conduct qualifying activities can claim the incentive retrospectively through their tax returns.

For AI companies, this program is particularly valuable given the inherently uncertain nature of artificial intelligence research and development. The program recognizes that AI advancement often involves significant technical challenges with no guaranteed outcomes, making it difficult for companies to secure traditional funding. By providing a 15% refundable tax credit on eligible R&D expenditure, the government effectively reduces the cost of AI innovation by sharing in the financial risk.

The program's objectives align closely with New Zealand's broader economic strategy to transition toward a knowledge-based economy. Key priorities include increasing business investment in R&D, encouraging companies to tackle complex technical challenges, supporting the development of intellectual property within New Zealand, and fostering collaboration between industry and research institutions. For the AI sector specifically, the program aims to position New Zealand as a competitive player in global AI markets while addressing local challenges through innovative technological solutions.

Recent policy refinements have clarified the treatment of cloud computing costs and software-as-a-service expenses, which are particularly relevant for AI companies. The program now explicitly recognizes that modern AI development relies heavily on cloud-based computing resources, machine learning platforms, and specialized software tools. This recognition ensures that AI companies can claim credits for the full spectrum of their development costs, from traditional salary expenses to cutting-edge cloud computing infrastructure.

The incentive operates without caps on claim amounts, making it particularly attractive for larger AI companies or those with substantial R&D programs. This uncapped structure recognizes that breakthrough AI innovations often require significant sustained investment over multiple years. Companies can claim the incentive for up to four previous tax years, providing flexibility for businesses that may not have been aware of the program or were still developing their R&D capabilities.

The refundable nature of the tax credit is crucial for many AI startups and early-stage companies that may not yet have significant tax liabilities. Rather than simply reducing tax owed, the program provides cash payments to eligible companies, effectively functioning as a government co-investment in their R&D activities. This feature makes the program accessible to companies at all stages of development, from pre-revenue startups to established enterprises expanding their AI capabilities.

Comprehensive Eligibility & Requirements

Understanding eligibility for the R&D Tax Incentive requires careful consideration of both the technical and administrative requirements. The program defines eligible R&D activities as systematic investigative or experimental activities that involve innovation or high levels of technical risk and are undertaken for the purpose of acquiring new knowledge or creating new or improved materials, products, devices, processes, or services.

For AI companies, eligible activities typically include developing novel machine learning algorithms that go beyond existing methodologies, creating AI applications to solve previously unsolved problems, advancing computer vision or natural language processing capabilities beyond current state-of-the-art, resolving significant technical uncertainties in AI implementation, and conducting fundamental research into AI architectures or training methodologies.

A critical aspect of eligibility is demonstrating genuine technical uncertainty. This means the R&D activities must involve challenges that cannot be resolved through routine engineering or by applying existing knowledge. For AI companies, technical uncertainty often manifests in questions such as: Will a new neural network architecture improve performance for a specific application? Can machine learning models be successfully adapted to work with limited training data in a particular domain? How can AI systems be made more interpretable while maintaining accuracy?

Common misconceptions about eligibility include the belief that all software development qualifies as R&D, that commercially available AI tools and platforms automatically make projects ineligible, or that successful outcomes are required to claim the incentive. In reality, routine software development, system integration using existing tools, and purely commercial activities typically do not qualify. However, using commercial AI platforms as part of genuine research into new applications or methodologies can still be eligible if the overall project involves systematic investigation and technical uncertainty.

Companies must maintain comprehensive documentation to support their claims. This includes detailed records of R&D activities, evidence of technical uncertainty and systematic approach, documentation of personnel time spent on eligible activities, records of all expenses related to R&D work, and evidence of the innovative nature of the work undertaken. The documentation should clearly demonstrate how the activities advance the company's technical knowledge and capabilities.

Pre-application preparation involves conducting a thorough assessment of R&D activities to identify eligible projects and expenses. Companies should establish robust time-tracking systems to capture personnel effort on R&D activities, implement project management processes that clearly distinguish R&D work from routine development, maintain detailed technical documentation of research approaches and findings, and establish expense tracking systems that can isolate R&D-related costs.

The eligibility assessment also considers the company's overall approach to R&D. Inland Revenue looks for evidence of systematic methodology, qualified personnel conducting the work, appropriate technical documentation, and clear connections between claimed activities and genuine innovation objectives. Companies should be prepared to demonstrate that their AI R&D activities represent genuine attempts to advance technical knowledge rather than routine application of existing techniques.

Funding Structure & Financial Details

The R&D Tax Incentive provides a 15% refundable tax credit on eligible R&D expenditure, with no upper limit on claim amounts. This uncapped structure is particularly beneficial for AI companies, which often require substantial investment in computing resources, specialized personnel, and extended development timelines. The refundable nature means companies receive cash payments even if they have no tax liability, making the program accessible to early-stage companies and startups.

Eligible expenses encompass a broad range of R&D-related costs. Salary and wage costs for employees directly engaged in R&D activities represent the largest category for most companies, including base salaries, bonuses, benefits, and employer contributions. Contractor and consultant fees for external parties conducting R&D work on behalf of the company are also eligible, provided the work meets the technical uncertainty and innovation criteria.

Cloud computing costs have become increasingly important for AI companies and are explicitly recognized under the program. This includes compute instances used for training machine learning models, storage costs for large datasets, specialized AI and machine learning platform fees, and data processing and analysis services. However, costs must be directly attributable to R&D activities rather than general business operations.

Materials and consumables used in R&D work qualify for the incentive, including specialized software licenses for development tools, datasets purchased for research purposes, hardware used exclusively or primarily for R&D, and testing materials and equipment. Depreciation on equipment used for R&D can also be claimed, though specific rules apply regarding the allocation of costs between R&D and non-R&D activities.

Certain expenses are explicitly excluded from eligibility. These include general business overheads not directly attributable to R&D, marketing and sales activities, routine maintenance and support, legal and patent costs (with limited exceptions), and capital expenditure on buildings and general-purpose equipment. Interest expenses and financing costs are also typically excluded.

Co-funding requirements are minimal—companies can combine the R&D Tax Incentive with other government grants and funding programs without penalty. This stackability makes the incentive particularly valuable as part of a broader funding strategy. However, companies must ensure they do not double-claim expenses across multiple programs and maintain clear records of all funding sources.

Payment timelines depend on when claims are filed and processed. Companies typically file claims with their annual tax returns, and Inland Revenue processes most claims within three to six months. For companies with no tax liability, refund payments are issued directly. The retrospective nature of the program means companies receive payments after completing R&D work and filing claims, rather than receiving upfront funding.

The program allows claims for up to four previous tax years, providing flexibility for companies that may not have previously claimed the incentive. This retrospective claiming can result in substantial cash injections for companies that have been conducting eligible R&D activities without claiming the incentive.

Application Process Deep Dive

The application process for the R&D Tax Incentive is integrated into New Zealand's standard tax return system, making it more straightforward than many traditional grant applications. However, the self-assessment nature of the program places significant responsibility on companies to correctly identify eligible activities and expenses while maintaining comprehensive supporting documentation.

The process begins with a thorough internal assessment of R&D activities conducted during the relevant tax year. Companies should systematically review all projects and activities to identify those that meet the eligibility criteria, focusing on work that involved technical uncertainty, systematic investigation, and genuine innovation. This assessment should result in a clear categorization of eligible versus ineligible activities, with detailed justifications for inclusion.

Documentation preparation is crucial and should be ongoing throughout the R&D process rather than assembled retrospectively. Key documents include project plans and technical specifications that demonstrate the innovative nature of the work, detailed time records showing personnel effort on eligible activities, expense records with clear attribution to R&D projects, technical reports and findings that evidence the systematic approach, and correspondence or meeting notes that demonstrate the technical challenges addressed.

When completing the tax return, companies must accurately report their R&D expenditure using the prescribed forms and categories. This involves calculating eligible salary and wage costs, including appropriate allocations for employees who split time between R&D and other activities, determining qualifying contractor and consultant expenses, identifying eligible material and equipment costs, and properly categorizing cloud computing and software expenses.

Common application pitfalls include over-claiming by including routine development activities that lack technical uncertainty, inadequate documentation to support claimed activities and expenses, incorrect allocation of shared costs between R&D and non-R&D activities, failure to properly distinguish between eligible and ineligible expenses, and insufficient technical detail to demonstrate the innovative nature of the work.

Evaluators from Inland Revenue assess claims based on several key factors. They examine whether claimed activities genuinely involve technical uncertainty and systematic investigation, review the competence and qualifications of personnel conducting R&D work, assess the adequacy of documentation and record-keeping, and verify the accuracy of expense calculations and allocations. Claims may be subject to audit, particularly for large amounts or first-time claimants.

To strengthen applications, companies should maintain detailed project documentation throughout the R&D process, implement robust time-tracking and expense allocation systems, engage qualified technical personnel to lead R&D activities, and clearly articulate the technical challenges and uncertainties addressed. Regular internal reviews of R&D activities can help ensure ongoing compliance with eligibility criteria.

Many companies benefit from engaging R&D tax specialists who understand the technical requirements and can help optimize claim values while ensuring compliance. These specialists can assist with eligibility assessments, documentation requirements, expense calculations, and ongoing compliance processes. While not mandatory, professional assistance often results in higher claim values and reduced audit risk.

Success Factors & Examples

Successful R&D Tax Incentive claims share several common characteristics that AI companies should emulate. The most critical success factor is demonstrating genuine technical uncertainty and systematic investigation. Successful AI projects typically involve novel approaches to well-defined technical challenges, with clear documentation of the experimental methodology and iterative development process.

Projects that have succeeded include developing proprietary machine learning algorithms for specific industry applications where existing solutions proved inadequate, creating AI systems that operate under unique constraints or requirements not addressed by commercial solutions, advancing the state-of-the-art in specialized AI domains such as computer vision for particular use cases, and conducting fundamental research into AI architectures or training methodologies with potential commercial applications.

For example, an AI company developing computer vision systems for agricultural applications might successfully claim R&D incentives for work on adapting deep learning models to operate effectively with limited training data specific to New Zealand farming conditions. The technical uncertainty lies in whether standard computer vision approaches can be successfully modified for this specialized application, and the systematic investigation involves experimenting with different model architectures, training approaches, and data augmentation techniques.

Another successful example might involve a company developing natural language processing capabilities for te reo Māori. The technical challenges include adapting existing NLP models to work with a language that has limited digital resources, developing appropriate training datasets, and creating culturally appropriate AI interactions. The systematic investigation would involve experimenting with transfer learning approaches, data collection methodologies, and model architectures specifically designed for low-resource languages.

Documentation quality significantly impacts claim success. Successful companies maintain detailed technical logs that capture the experimental process, decision points, and findings. They clearly articulate the technical problems being addressed and why existing solutions are inadequate. Project documentation demonstrates the systematic approach taken and shows how the work advances the company's technical capabilities.

Common reasons for rejection include claims that describe routine software development rather than genuine R&D, insufficient documentation to support the technical uncertainty and systematic investigation claims, over-allocation of expenses to R&D activities that actually involve routine work, and failure to demonstrate that the work represents a genuine advance in technical knowledge or capability.

Companies should focus on demonstrating impact and return on investment by clearly articulating the technical advances achieved through their R&D work, showing how the research contributes to the company's competitive position, documenting any intellectual property or proprietary knowledge developed, and measuring improvements in technical capabilities or performance metrics.

The most successful AI companies treat the R&D Tax Incentive as part of a broader innovation strategy rather than simply a compliance exercise. They use the documentation and assessment requirements as opportunities to systematically evaluate their R&D processes, identify areas for improvement, and demonstrate value to stakeholders including investors and partners.

Strategic Considerations

The R&D Tax Incentive should be viewed as one component of a comprehensive funding strategy rather than a standalone solution. Its retrospective nature and broad eligibility criteria make it highly complementary to other funding programs, particularly those that provide upfront capital or target specific development stages.

Companies often combine the R&D Tax Incentive with Callaghan Innovation grants, which provide upfront funding for specific innovation projects. The timing works well—companies can use grant funding to conduct R&D activities and then claim tax incentives on the eligible expenses. Similarly, the incentive complements venture capital and angel investment by effectively reducing the cost of R&D activities and improving overall project economics.

The decision of when to apply for the R&D Tax Incentive versus alternatives depends largely on timing and funding needs. Companies requiring upfront capital should prioritize grants and investment, while those with sufficient capital to conduct R&D activities can rely on the tax incentive for retrospective cost recovery. The incentive is particularly valuable for companies with ongoing R&D programs that can benefit from the predictable, uncapped nature of the support.

Post-award compliance involves maintaining detailed records and being prepared for potential audits. Inland Revenue may review claims, particularly for large amounts or companies making their first claims. Companies should maintain all supporting documentation for the required retention period and be prepared to provide additional information if requested. Regular internal reviews of R&D activities and expense allocations help ensure ongoing compliance.

Relationship management with Inland Revenue is important for companies making regular claims. This involves responding promptly to any requests for information, maintaining accurate and comprehensive records, seeking clarification when eligibility questions arise, and potentially engaging with IRD's R&D tax specialist teams for guidance on complex technical issues.

The long-term strategic value of the R&D Tax Incentive extends beyond immediate financial benefits. Regular claiming helps establish a track record of innovation activity that can be valuable when seeking additional funding or partnerships. The documentation requirements encourage systematic approaches to R&D that often improve project outcomes and knowledge management.

Companies should also consider the international implications of their R&D activities. New Zealand's R&D Tax Incentive can be combined with similar programs in other jurisdictions, though careful attention to tax treaties and transfer pricing rules is required for companies with international operations.

The program's uncapped nature makes it particularly valuable for companies planning significant scaling of their R&D activities. Unlike grants with fixed amounts, the tax incentive grows with the company's R&D investment, providing ongoing support for expansion and increased innovation activity.

For AI companies specifically, the program's recognition of cloud computing costs and modern development methodologies makes it well-suited to supporting the full lifecycle of AI product development, from initial research through to deployment and iteration. This comprehensive coverage, combined with the program's flexibility and accessibility, positions the R&D Tax Incentive as a cornerstone of funding strategy for New Zealand AI companies committed to genuine innovation and technical advancement.