The Statutory Architecture of the Research Credit and the IUS Exemption

The Research and Development tax credit was introduced in 1981 to combat a perceived decline in American industrial innovation. However, the initial broadness of the credit led to significant controversy regarding what activities truly constituted “qualified research”. By 1986, Congress introduced the “Four-Part Test” to ensure that the credit was applied only to activities that were technological in nature and involved a structured process of experimentation intended to eliminate technical uncertainty.

The Emergence of the Internal-Use Software Conflict

Section 41(d)(4)(E) generally excludes research on software developed primarily for internal use, except to the extent provided by regulation and subject to statutory exceptions for software used in qualifying research or qualifying production processes. The statute does not itself set out a high-threshold-of-innovation test. Treasury regulations supply that additional test for internal-use software to which it applies.

The definition of what constitutes IUS has shifted significantly across different regulatory regimes. The primary friction point between the IRS and taxpayers has been the boundary between software used for “General and Administrative” (G&A) functions and software that enables interaction with third parties.

FedEx Corp. v. United States (2009): The Rejection of the Discovery Test

In FedEx Corp. v. United States, No. 08-2423 (W.D. Tenn. June 9, 2009), FedEx sought partial summary judgment on the legal standards governing research-credit claims for 1997 through 2000. It had begun developing a new billing and revenue-control system in 1996 and abandoned the project in 2001 after incurring substantial development costs. The dispute concerned which discovery and internal-use-software rules governed those historical claims. The order resolved the applicable standards, leaving factual qualification to be established.

The Discovery Test Controversy

The 2001 final regulations imposed a discovery standard based on advancement beyond the knowledge ordinarily available to skilled professionals. The 2003 final regulations rejected that standard in favor of resolving technical uncertainty. The government relied on Announcement 2004-9 to argue that taxpayers using the 2001 internal-use-software provisions also had to accept the 2001 discovery standard. FedEx argued that it could use the 2001 software provisions without that superseded discovery requirement.

The court granted FedEx’s motion for partial summary judgment. It permitted reliance on the 2001 internal-use-software test without the 2001 discovery test and rejected the government’s attempt to impose the older discovery requirement through Announcement 2004-9. The decision addressed the interaction of successive regulatory regimes for the years before the court. It did not originate the uncertainty standard, which Treasury had already adopted, or award a blanket credit for the project.

The High Threshold of Innovation Test

The 2009 order allowed FedEx to rely on the three-part high-threshold-of-innovation test in the 2001 final regulations. Its principal elements were innovation, significant economic risk, and lack of commercial availability. Modern claims must use the regulations applicable to their tax years; the 2016 final regulations revised and clarified the internal-use-software framework. The historical test considered in FedEx included:

Innovation: The software must be intended to result in a reduction in cost, improvement in speed, or other improvement that is substantial and economically significant.

Significant Economic Risk: The taxpayer must commit substantial resources to the development, and there must be substantial uncertainty, because of technical risk, that such resources would be recovered within a reasonable period.

Commercial Availability: The software cannot be purchased, leased, or licensed and used for the intended purpose without modifications that would satisfy the first two requirements.

For claimants facing the same historical regulatory issue, the order supported a more favorable discovery standard while retaining the additional internal-use-software requirements. Large investment, commercial importance, project failure, or use in logistics or financial services does not independently establish credit eligibility.

Technical and Evidentiary Requirements for Qualified Research

The FedEx precedent did not remove the necessity of the “Four-Part Test” found in Section 41(d). For a project to qualify, it must meet all four criteria, even if it passes the HTI test. The contemporary application of these tests, particularly in the wake of Phoenix Design Group v. Commissioner (2024), reveals a critical need for granular, activity-level documentation.

The Four-Part Test Under Modern Scrutiny

The interaction between statutory language and judicial interpretation has refined the requirements for each prong of the qualified research test:

Permitted Purpose: The research must concern a new or improved business component and relate to its function, performance, reliability, or quality. Changes directed only to style, taste, cosmetic, or seasonal design factors do not meet this requirement.

Technological in Nature: The process of experimentation must fundamentally rely on principles of physical or biological sciences, engineering, or computer science.

Elimination of Uncertainty: The research must seek to resolve uncertainty about the capability, method, or appropriate design of the component.

Process of Experimentation: Substantially all of the relevant research activities must constitute elements of an evaluative process addressing technical uncertainty for a qualified purpose. Modeling, simulation, and systematic trial and error may provide evidence; routine testing or a description of the final solution alone is insufficient.

The Impact of Loper Bright on IRS Rulemaking and R&D Litigation

Loper Bright Enterprises v. Raimondo, decided by the Supreme Court on June 28, 2024, changed the framework for reviewing agency interpretations of statutes. Its application in the separate FedEx foreign-tax-credit litigation is relevant by analogy to tax-regulation disputes, but neither case supplies a new substantive test for research-credit eligibility.

The End of Chevron Deference

Under Chevron, courts generally deferred to reasonable agency interpretations when the governing statute was ambiguous and the doctrine’s other conditions were met. Loper Bright overruled that framework and requires independent judicial interpretation. Agency expertise can still be persuasive, and courts must respect authority that Congress actually delegated. The decision did not automatically invalidate existing tax regulations or overturn earlier judgments sustaining regulations.

The FedEx district court invalidated the disputed foreign-tax-credit regulation in 2023 and rejected the government’s renewed argument on February 13, 2025, after considering Loper Bright. The government appealed in August 2025; FedEx’s fiscal 2026 annual filing describes that appeal and the associated uncertainty. These rulings concern foreign-tax credits associated with transition-tax earnings, not Section 41. Their reasoning may inform other challenges, but they do not authorize taxpayers simply to disregard research-credit regulations.

Implications for Section 41 Challenges

Independent statutory review may support a properly grounded challenge to a particular regulation. The following are potential lines of legal analysis, not holdings that any Section 41 rule has been invalidated. An argument must account for statutory language, delegated authority, controlling precedent, jurisdiction, and the regulation applicable to the claim year.

Reviewing the Shrinking-Back Rule: This rule permits testing progressively smaller components when the larger business component fails the qualified-research requirements. Any challenge must distinguish that regulatory mechanism from the substantially-all requirement, which appears in Section 41 itself. Neither Loper Bright nor the FedEx rulings discussed here eliminated either requirement.

Reviewing the Definition of Process of Experimentation: A taxpayer may contest a particular regulatory interpretation through an appropriate proceeding, but the statute expressly requires experimentation. Standard industrial testing is not automatically eligible following Loper Bright; the taxpayer must establish the relevant uncertainty and actual evaluative activities.

Reviewing Internal-Use-Software Definitions: Section 41(d)(4)(E) expressly contemplates regulations governing the exclusion. A challenge must therefore address the scope of that delegation. The 2016 general-and-administrative and third-party-interaction rules remain relevant to applicable years; software classification cannot be replaced by a taxpayer’s preferred interpretation without legal support.

Funded Research and State Law: The System Technologies Precedent

Another critical area of future litigation involves the “Funded Research” exclusion under Section 41(d)(4)(H). The IRS frequently attempts to deny credits to contractors, arguing that their research is funded by their customers.

In System Technologies, Inc. v. Commissioner, Docket No. 12211-21 (order dated January 3, 2025), the Tax Court denied the Commissioner’s motion for partial summary judgment on funded research. Indiana law governing the contracts allowed customers to recover payments if the promised systems were not delivered. The court therefore concluded that payment depended on successful research for the issue presented. This interlocutory order did not determine that every claimed activity or expense qualified. It illustrates why the contracts and governing state-law remedies must be analyzed together.

Quantitative Modeling of the R&D Tax Credit

The regular research credit and the alternative simplified credit (ASC) use different historical comparisons. Eligibility, elections, aggregation, and the rules applicable to the tax year affect the computation; the following formulas describe the ordinary incremental QRE component before applicable Section 280C adjustments and tax-liability limitations.

The Credit Calculation Formula

The regular credit’s ordinary incremental QRE component is calculated as:

Regular credit = 20% × max(0, current-year QREs − base amount).

The regular-credit base amount generally incorporates both a gross-receipts calculation and a minimum base:

Base amount = greater of (fixed-base percentage × average annual gross receipts for the four preceding tax years) or (50% × current-year QREs). Special fixed-base-percentage rules can apply. The ASC generally equals 14% × max(0, current-year QREs − 50% of average QREs for the three preceding tax years). If any of those three years has no QREs, the ASC generally equals 6% of current-year QREs.

The June 9, 2009 FedEx order addressed discovery and internal-use-software standards; it should not be cited as a ruling requiring intercompany receipts in the credit base. Treasury Regulation Section 1.41-6(i) generally disregards transactions between members of the same group for credit computation. Gross receipts and aggregation must instead be determined under the applicable Section 41 rules, with attention to group membership and consistent treatment across years.

Sections 174, 174A, and 41

For tax years beginning in 2022 through 2024, the general Section 174 regime required five-year amortization of domestic research expenditures and fifteen-year amortization of foreign research expenditures. Public Law 119-21, enacted July 4, 2025, added Section 174A, generally restoring current deductions for domestic research or experimental expenditures for tax years beginning after December 31, 2024. Alternative capitalization and transition provisions apply, including options concerning unamortized domestic costs from 2022–2024. Foreign research expenditures generally remain subject to fifteen-year amortization under Section 174. Deductibility does not establish credit eligibility: Section 41 imposes additional tests and exclusions, and Section 280C coordinates the credit with deductions. Phoenix Design Group applied the research-expense and experimentation requirements for its earlier claim years; it was not a ruling on the 2025 legislation.

The Future of Documentation and Audit Defense

The useful lesson from these cases is to distinguish questions of legal interpretation from proof of qualifying activity and expense. Loper Bright changed statutory review, but did not shift all factual burdens to the IRS or remove the ordinary substantiation requirements applicable to research-credit claimants.

Lessons from Phoenix Design Group (2024)

In Phoenix Design Group, Inc. v. Commissioner, T.C. Memo. 2024-113, the Tax Court found that none of three trial projects qualified, including at the shrinking-back level. The case involved research years 2013–2016 and stipulated consequences for penalties. Its core problem was failure to establish the required uncertainty and investigative or experimental work through the evidence presented, not a standalone rule forbidding retrospective studies.

The opinion highlights three connected evidentiary weaknesses:

General Design Process: Describing a six-stage engineering workflow did not establish the specific investigative activities or experimentation performed on the trial projects. The possibility of later revisions did not itself prove continuing uncertainty over the entire system.

Insufficie