Financial services institutions have long invested in software and technology to facilitate business operations and serve customers. In recent years, that investment has increasingly incorporated artificial intelligence (AI) in more meaningful ways.
These AI investments are helping companies develop software, automate processes, analyze data, and solve complex technical problems. AI is also influencing the pace of development, the cross-functional teams involved, and the mix of resources required to support these efforts.
The continued adoption of AI, combined with recent legislative and administrative changes related to the tax treatment of research and development (R&D) costs, has created a dynamic environment that should be considered as part of this year’s tax filing. These considerations include both compliance-related risks and opportunities to increase incentives associated with R&D spending, which include turning investments in AI into tax savings.
This article will focus on the following 10 recommendations for financial institutions to consider in regard to R&D spending and associated AI investments.
1. Understand the opportunity for tax incentives.
The R&D tax credit was introduced in 1981 to incentivize investment in R&D activities through a reduction in tax liability. The credit provides a dollar-for-dollar offset of tax liabilities, subject to certain limitations.
The credit is available for qualified research activities related to the development or improvement of a product, process, computer software, technique, formula, or invention. These activities must be intended to achieve new or improved functionality, performance, reliability, or quality. Qualification does not require successful development, nor does it require creating entirely new knowledge within an industry, field of science, or engineering.
The credit is calculated by comparing current-year spending to a historical base amount, with applicable credit rates generally ranging from 6% to 20%, depending on the calculation method used. The R&D tax credit can be carried back one year and forward 20 years in situations where there is not sufficient tax liability to utilize the credit in the year in which the credit is generated. Many states also provide tax credits for R&D activities.
2. Evaluate software development activities and investments.
Consider how your organization has invested in software and technology in recent years. Have these investments resulted in new or improved functionality, performance, reliability, or quality within your software platforms or systems? In addition, assess whether these investments represent purchased software or costs associated with developing a software solution under the Internal Revenue Code. If spending is classified as software development that creates new or improved features, functionality, or capabilities, there may be an opportunity to treat these costs as qualified research expenses (QREs) for tax purposes and claim R&D tax credits to help offset the cost of these investments.
3. Understand your software development life cycle (SDLC).
Developing an understanding of your organization’s SDLC can provide insight into how related activities and costs should be reflected in your tax returns. Key considerations include how the stages of the SDLC align with tax definitions of R&D activities, who within the organization participates in each stage of development, and what documentation or project/cost tracking exists to support these activities. AI’s increasing involvement is also changing the mix of personnel participating in development and the activities they perform. Understanding these changes may help identify a broader cross-section of employees and roles whose activities could be considered in your R&D tax credit.
4. Assess how development dollars are being spent.
In addition to changing the teams that participate in the development life cycle, AI is also influencing the types of costs incurred to support development efforts. These costs may include investments in infrastructure, cloud hosting, AI licensing, compute usage, e.g., tokens, and third-party consulting. Developing an understanding of how these costs are incurred and categorized is important for evaluating how they should be treated in your tax return and for enhancing available deductions and credits.
5. Evaluate the impact of AI on the pace of development.
AI is accelerating development across SDLC phases—including requirements and design, build, testing, and deployment—by enabling faster translation of business requirements into technical specifications, improving issue triage and test generation, and enhancing security and compliance reviews through automated policy and vulnerability checks.
Understanding how this acceleration affects development costs within a given period, as well as how the expanded use of AI and increased automation aligns with tax definitions of R&D, is important for projecting the timing of deductions and potential tax incentives. Companies should also evaluate how AI is being incorporated into the SDLC to assess whether underlying activities align with R&D tax credit qualification requirements.
6. Assess where development activities are performed.
The location of development activities can significantly impact tax treatment. Onshore development activities may qualify for immediate expensing and may be eligible for the R&D tax credit, while offshore development activities are generally required to be capitalized and amortized over a 15-year period and are not eligible for the R&D tax credit.
For onshore activities, the state in which development occurs can also affect tax treatment. Some states have not conformed to the One Big Beautiful Bill Act’s (OB3’s) provisions related to domestic research and experimental expenditures (DREEs), which may impact the timing of deductions for R&D spending when determining state tax liability. In addition, certain states offer their own R&D tax credits for development activities conducted within their jurisdiction, which may help maximize the tax benefits of development investments.
7. Consider your options under the OB3 and make your elections.
Under the OB3, companies have options for the federal treatment of DREEs for tax years beginning after December 31, 2024, including the ability to immediately expense such costs. Companies may also elect to accelerate the amortization of DREEs that were previously capitalized under Tax Cuts and Jobs Act (TCJA) guidance.
Modeling the impact of these options and elections is an important step in determining the most effective approach, including evaluating how accelerated deductions may affect taxable income and other tax concepts and attributes. It is also important to understand the process of making these elections. Depending on the approach, companies may need to file a statement in lieu of Form 3115 to adopt a method under the OB3 or make an election under the applicable transition guidance for costs previously capitalized under the TCJA. In addition, the acceleration of the unamortized tax basis of DREEs in tax years beginning after December 31, 2024 may trigger a tax loss, which, in turn, may require any R&D tax credit claimed in 2025 to first be carried back to the prior tax year, requiring amended return claims and documentation to be compiled in support of Chief Counsel Memorandum 202114101F.
8. Evaluate the impact of IRC §280C.
Internal Revenue Code (IRC) Section 280C provides guidance on the treatment of expenses used to support a tax credit. When R&D costs are included in the R&D tax credit, §280C generally requires a corresponding reduction to the deduction associated with those costs by the amount of the credit claimed.
Companies can account for this impact either through an election on the R&D tax credit form or by making a book-to-tax adjustment to allowable deductions. The application of IRC §280C was affected by the requirement to capitalize DREEs and the conforming amendment under the TCJA, and many companies interpreted these changes to limit the adjustment in certain situations. As a result, some companies were able to claim larger R&D tax credits in recent years without reducing the credit or related deductions.
However, §280C was modified under the OB3, reinstating its impact on R&D deductions. Companies should reassess how this provision applies to their R&D tax credit calculations and related deductions, particularly in situations where they are in a net operating loss position.
9. Prepare for evolving R&D tax credit reporting requirements.
In recent years, the IRS has issued guidance in multiple forms to indicate its intent to obtain more detailed information with R&D tax credit claims, primarily through updates to Form 6765 and its instructions. Beginning with 2024 filings, companies were required to provide additional disclosures related to their R&D tax credit claims, including a count of their business components and other informational items that could impact credit calculations.
Additional reporting requirements are scheduled to take effect for 2026 credit claims and may be more consequential. Companies will be required to provide detailed information on each business component representing 80% of total qualified research expenses, up to a maximum of 50 business components. This includes listing each business component, identifying its type, and reporting the amount of each category of QREs associated with it.
Companies should begin planning now to help ensure they can capture and report the level of detail required to comply with these expanded reporting requirements.
10. Assess potential for an uncertain tax position.
Companies should evaluate whether a reserve for uncertain tax positions is appropriate in connection with their R&D tax credit. It is not uncommon for companies to reserve a portion of the credit based on the specific facts and circumstances of their claim.
Financial service institutions should use the information above to reflect on activities performed in the current tax year, including AI-related development, and proactively plan for future investments. This includes incorporating these considerations into decisions regarding where investments are made, how costs are categorized and captured, and how documentation is maintained.
Taking a proactive approach can help companies better project taxable income, enhance tax incentives associated with R&D spending, and meet applicable compliance and reporting requirements related to their DREE deductions and R&D tax credit claims. If you have any questions or need assistance, please reach out to a professional at Forvis Mazars.