Saykally v. Commissioner and the Evolving Federal R&D Tax Credit Standard under Sections 41 and 174

The architectural framework of the United States tax system has long utilized targeted incentives to catalyze technological innovation, primarily through the Research and Development tax credit under Section 41 and the deduction of research and experimental expenditures under Section 174 of the Internal Revenue Code. These provisions serve as the cornerstone of federal support for the private sector’s investment in technical advancement, yet their application remains one of the most litigated and complex areas of tax law. The case of Saykally v. Commissioner represents a fundamental touchstone in this legal landscape, particularly regarding the threshold requirement that research be conducted in connection with a trade or business. As the Internal Revenue Service (IRS) and the federal courts have transitioned from a historically lenient “discovery” standard to a rigorous, documentation-heavy “process of experimentation” standard, the precedents established in Saykally and subsequent cases such as Little Sandy Coal Co. v. Commissioner and Phoenix Design Group, Inc. v. Commissioner have redefined the risk profile for taxpayers claiming these incentives. See also the difference between the Section 41 credit and Section 174 expenses.

The Jurisprudential Foundation of Section 174 and the Trade or Business Requirement

The case of Saykally v. Commissioner, primarily documented in T.C. Memo. 2003-152 and later affirmed in an unpublished per curiam decision by the Ninth Circuit (247 F. Appx. 914), centers on the eligibility of software development costs for immediate deduction under Section 174. David M. Saykally, a veteran of the computer software industry, sought to deduct approximately $1.4 million in research and development expenses for the taxable years 1995 and 1996. The expenditures were incurred through CPSG, Inc., and its subsidiaries, which paid the costs on behalf of the petitioner. The core legal dispute was whether these activities were undertaken “in connection with a trade or business,” a prerequisite for the Section 174 deduction.

The facts of the case revealed that while Saykally had significant technical expertise and had indeed conducted research, he lacked the objective intent to enter into a future business of his own to commercialize the resulting technology. Instead, the evidence indicated an intent to license the technology to existing businesses or to conduct business through other entities rather than establishing an independent trade or business. This distinction is critical because Section 174 requires that the taxpayer be the one intended to carry out the business activities related to the research.

The Tax Court’s finding, and the Ninth Circuit’s subsequent affirmation, emphasized that a taxpayer must demonstrate a realistic prospect of entering into a trade or business to qualify for the deduction. In the software context, this has profound implications. If a developer creates software for a third party who then markets and sells that software, the developer may fail the trade or business test unless they maintain an active, entrepreneurial stake in the commercial outcome. The Saykally ruling serves as a cautionary tale for independent contractors and technical founders who may focus on the technical merits of their innovation while neglecting the structural and intent-based requirements of the tax code. Pre-revenue founders should also review whether they qualify for R&D tax credits as a pre-revenue startup.

Evolution of the Four-Part Test and the Process of Experimentation

While Saykally addressed the threshold of business intent, the technical qualification of research activities is governed by the rigorous “four-part test” established under Section 41(d). This test ensures that the credit is reserved for activities that are technological in nature and involve a systematic process of experimentation to resolve technical uncertainty. For a plain-language walkthrough, see how the four-part test works in simple terms.

The Permitted Purpose and Section 174 Tests

The first two prongs of the test require that the research be for a permitted purpose—developing a new or improved product or process—and that it be undertaken to eliminate uncertainty. Uncertainty exists if the information available to the taxpayer does not establish the capability or method for developing the product, or the appropriate design. Litigation such as Phoenix Design Group, Inc. v. Commissioner (T.C. Memo. 2024-113, decided December 23, 2024) has significantly raised the bar for what constitutes “uncertainty.”

In Phoenix Design Group, the taxpayer, a multidisciplinary engineering firm, argued that its engineering calculations for air handling, plumbing, and fire protection systems (MEPF systems) were necessary to resolve technical uncertainties. The court rejected this, stating that performing routine calculations on available data is not an investigative activity because the taxpayer already possesses the information necessary to solve the unknown. This ruling suggests that for future applications, taxpayers must clearly identify the specific technical information that was missing at the start of the project and document the investigative steps taken to acquire it. Architecture and engineering practices should compare this holding with R&D tax credits for architecture and engineering firms and whether architects can claim R&D tax credits.

The Technological Nature and Process of Experimentation Prongs

The third prong requires the research to be technological in nature, meaning it must fundamentally rely on principles of physical or biological sciences, engineering, or computer science. Finally, the fourth prong, the “Process of Experimentation,” requires that substantially all of the activities constitute a process of evaluating alternatives through modeling, simulation, or systematic trial and error.

The court in Phoenix Design Group found that the firm’s six-stage design process was largely linear rather than genuinely iterative, and that its time-tracking narratives did not align with what was supposed to occur at each stage. Simply communicating results to an architect or building owner was not considered a process that mirrors the scientific method. This highlights a growing trend where the IRS and courts look for hypothesis-driven testing rather than standard professional design cycles. For future R&D applications, the implication is that taxpayers must maintain logs of failed attempts, rejected alternatives, and the specific technical reasons why a particular design was chosen over others. Practical documentation methods are covered in how to document a process of experimentation for tax purposes and how to track R&D hours for tax credits.

The “Substantially All” Requirement and the Mathematics of Compliance

A critical component of the R&D credit is the requirement that “substantially all”—interpreted as 80 percent or more—of the research activities for a business component must constitute elements of a process of experimentation. If this threshold is not met at the whole-project level, the taxpayer may apply the “shrink-back rule,” which allows the test to be applied at a subcomponent level. See how to calculate the substantially all rule for R&D.

Lessons from Little Sandy Coal Co. v. Commissioner

The Seventh Circuit’s decision in Little Sandy Coal Co. v. Commissioner (2023) provided a comprehensive analysis of how to calculate this 80 percent threshold. The case involved Corn Island Shipyard, a shipbuilding subsidiary of Little Sandy Coal, which constructed 11 first-in-class vessels in a tax year ending in June 2014. The taxpayer argued that because each vessel was substantially new, the entire project should be treated as research. The court rejected this “novelty argument,” ruling that the substantially all test applies to activities (measured on a cost or other consistently applied reasonable basis) rather than the physical elements or novelty of the product. Manufacturers evaluating process work should also review whether manufacturing process improvements qualify for R&D credits.

The court utilized a fraction to determine eligibility, where the numerator included only activities that were elements of experimentation and the denominator included all activities related to the business component. A significant issue arose regarding the classification of production workers, direct supervision, and direct support wages. The Seventh Circuit clarified that expenses properly includable as qualified research expenses under Section 174—including direct support and direct supervision—should be incorporated in both the numerator and denominator of the fraction, and it left open whether pilot-model production activities could themselves constitute elements of a process of experimentation. Ultimately, however, the taxpayer failed to offer a “principled way” to determine what portion of employee activities constituted experimentation, and the Seventh Circuit affirmed the Tax Court’s disallowance of the credit for all 11 vessels.

Funded Research and the Allocation of Financial Risk

Section 41(d)(4)(H) excludes “funded research” from the credit, meaning research where the taxpayer does not bear the financial risk of failure or does not retain substantial rights to the results. For firms operating under contracts, this has historically been a high-stakes battleground with the IRS. Related questions include whether grant funding disqualifies an R&D tax credit claim and whether outsourced contractor development can still qualify.

Contractor Funding Disputes: Smith and System Technologies

In early 2025, the Tax Court denied the IRS’s motions for summary judgment in both Smith v. Commissioner and System Technologies, Inc. v. Commissioner, allowing both cases to proceed to trial on the “funded research” question. In the Smith litigation, the taxpayers were partners of Adrian Smith + Gordon Gill Architecture, LLP (AS+GG), an architecture firm known for designing large-scale “supertall” towers, and the dispute centered on research credits claimed for a sample of six international projects for tax years 2008 through 2010. The IRS argued the research was funded because the client contracts were not, in its view, structured to place financial risk on the firm.

The case ultimately went to trial, and the Tax Court issued its final decision in June 2026 (T.C. Memo. 2026-50). The result was mixed rather than a clean taxpayer win: the court found that none of the six sampled contracts were “contingent on the success” of the research—because payment was tied to percentage-of-completion or hourly billing rather than technical success—meaning the research was funded under that first prong of the regulatory test. However, because AS+GG retained substantial rights in four of the six projects (discussed below), the firm was still entitled to a partial research credit for those projects. The court also ruled in the taxpayers’ favor on a separate issue, holding that the partners’ compensation was reasonable under Section 174(e) using the Seventh Circuit’s “independent investor” test. The System Technologies case, involving a similar funded-research dispute governed in part by state contract law, survived the IRS’s summary judgment motion on similar reasoning, though a final trial outcome was not identified as part of this review.

Substantial Rights and Intellectual Property

The “rights” prong of the funded research test requires that the taxpayer retain substantial rights to the research results, even if those rights are shared with the client. In the Smith litigation, the Tax Court’s final decision examined each of the six sample contracts individually. For two of the projects, the contracts vested exclusive property rights and copyrights entirely with the client, and the court held that needing to secure the client’s permission to use the research—with no limit on the client’s ability to withhold that consent—meant AS+GG did not retain substantial rights in those two projects. For the remaining four projects, however, AS+GG had negotiated licenses or retained copyrights that allowed it to reuse and market the underlying research, and the court confirmed that the right to use research results, even on a non-exclusive basis, is itself a substantial right. The court also declined to let foreign law (such as the law of the United Arab Emirates or the United Kingdom) override the express terms of the contracts, holding that the funding analysis must be based on the research agreement itself. This is a meaningful, if qualified, win for engineering and design firms: transferring a final deliverable to a client does not automatically disqualify a firm from the credit, but the specific rights retained in each contract will determine the outcome project by project.

The Evolving Standard of Contemporaneous Documentation

Across the analyzed cases—from the “trade or business” failure in Saykally to the “routine engineering” failure in Phoenix Design Group—a common thread is the failure of documentation. The IRS and the courts are no longer accepting “look-back” studies or arbitrary estimates of time spent on R&D. Claimants should maintain contemporaneous documentation rather than reconstructed narratives. See what constitutes a contemporaneous record for R&D tax credits and what documents are needed for an R&D tax credit study.

The High Bar for Activity-Level Records

The Phoenix Design Group ruling is particularly illustrative of the consequences of poor record-keeping, as the court upheld a 20% accuracy-related penalty (stipulated by the parties) because the taxpayer lacked activity-level records linking employee time to the specific four-part test criteria. For future applications, the implication is that businesses must implement systems to document:

1. Technical uncertainties at the project’s inception.
2. Hypotheses being tested.
3. The specific engineering or scientific principles applied.
4. Detailed results of iterative tests and alternative evaluations.

Timekeeping questions frequently arise in this setting, including whether timesheets are required for R&D tax credits and whether a taxpayer can reconstruct R&D time estimates if hours were not tracked.

The Shrink-Back Rule as a Tactical Documentation Strategy

When a project as a whole fails the 80% test, the shrink-back rule can save a portion of the credit, but only if the documentation supports the subcomponent level. In Phoenix Design Group, the court applied the shrink-back rule to subsets of the MEPF systems at issue but found that none of those subsets qualified either, because the record lacked the necessary detail