The IRS Is Attacking R&D Credits — And They May Be Doing It Illegally

Did a tax firm promise you a research and development credit you didn't know you had? Here's what's actually happening, what the IRS gets wrong about these credits, and why a landmark Supreme Court decision may change the entire game.

What the IRS Is Doing Right Now

We know from inside sources that IRS Appeals officers are currently in training on one specific topic: examination of research credits under IRC Section 41.

That detail matters. When Appeals officers receive specialized training on a credit, it means examination-level disputes on that credit are landing at Appeals in volume. The IRS doesn't train Appeals officers on issues that aren't generating significant controversy. Something is happening.

What's happening is this: over the past several years, a cottage industry of R&D credit promoters emerged, cold-calling small and mid-size businesses — restaurants, contractors, manufacturers, farmers — telling them they qualified for substantial research and development tax credits they never knew about. These firms charged contingency fees of 20 to 30 percent of the credit calculated, produced thick binders of documentation, and filed amended returns claiming refunds.

Now the IRS is pushing back. Hard.

If you received a research credit study from a third-party firm and claimed credits on an amended return, you may be in the IRS's crosshairs. And if you're already in examination or heading to Appeals, you need to understand three things the compliance industry won't tell you.

First: What IRC Section 41 Actually Says

The research credit under IRC Section 41 was created in 1981 to encourage American companies to invest in research and development. It provides a credit of up to 20 percent of qualified research expenses above a base amount.

To qualify, research must meet a four-part test: it must be technological in nature, serve a qualified purpose, involve the elimination of uncertainty, and employ a process of experimentation.

Notice what the statute does not require. It does not require that research succeed. It does not require a patentable invention. It does not require a commercially viable product. It requires a process undertaken to resolve genuine uncertainty.

The IRS, through its regulations, has interpreted these requirements so narrowly that they effectively transform the Research and Development credit into a Development-only credit. And that interpretation may be illegal.

The IRS Forgot About the "R" in R&D

Here is the argument that no compliance-oriented CPA will ever raise with you, because their entire practice assumes IRS regulatory definitions are correct.

The credit is called the Credit for Increasing Research Activities. The word "research" is right there in the statute. Research, in its ordinary meaning, includes investigation that produces negative results. Dead ends. Failed experiments. Studies that conclude a particular approach won't work.

Think about how actual research works. A pharmaceutical company funding early-stage drug discovery doesn't know which compounds will advance to clinical trials. Most won't. The research that eliminates dead ends is scientifically essential — and economically valuable. The company now knows what doesn't work. That knowledge has value. Bell Labs produced some of the most important technological breakthroughs of the 20th century precisely by funding research that produced negative results alongside positive ones.

Now consider a practical example. Suppose a company wants to build a sophisticated AI tool to analyze patterns in IRS examination data. The research phase involves reading IRS publications, studying audit selection methodologies, interviewing tax professionals, reviewing publicly available data, and investigating which analytical approaches are feasible. Much of that investigation produces dead ends. Certain approaches don't work. Certain data sources turn out to be insufficient.

The IRS would say that reading publications and investigating dead ends doesn't qualify as research under Section 41. But the statute says "research activities." Dead-end research is research. It resolved uncertainty — in the negative direction.

The IRS has read a success requirement into a statute that contains no such requirement. The regulation is narrower than the statute permits.

Why This Matters More Than Ever: The Loper Bright Revolution

In 2024, the Supreme Court decided Loper Bright Enterprises v. Raimondo, overruling the Chevron doctrine that had governed administrative law for forty years.

Under Chevron, courts deferred to agency regulatory interpretations of ambiguous statutes. If the IRS said "research" meant only successful development-oriented activity, courts largely accepted that because Treasury had adopted regulations saying so.

Loper Bright changed everything. Courts must now independently determine what a statute means. They are no longer permitted to simply defer to the agency's preferred interpretation. If the statute is ambiguous — and the word "research" is broad enough to be genuinely ambiguous — the court must determine the best interpretation of that word, not the IRS's preferred interpretation.

This is enormously significant for Section 41 cases. The question of whether "research activities" includes genuine investigative work that produces negative results, or only development-oriented activity that produces technologically novel results, is now an open legal question. Courts must answer it independently.

We believe the honest answer — the one a court must now reach without deference to Treasury — is that "research activities" in the ordinary meaning of the word includes investigation that resolves uncertainty regardless of outcome. The IRS's regulatory framework, which effectively requires successful development, is narrower than the statute permits.

Every Section 41 disallowance that rests on the IRS's regulatory definition of qualified research is now potentially challengeable on Loper Bright grounds.

The Penalty Problem — And the Jarkesy Angle

If you're facing a Section 41 examination, the IRS isn't just threatening to disallow the credit. They're threatening penalties.

The IRS's Audit Techniques Guide specifically instructs examiners to evaluate the Section 6676 erroneous claim for refund penalty, which can be 20 percent of the disallowed amount, and the accuracy-related penalty under Section 6662. On a substantial research credit claim, these penalties can be significant.

Here's what the compliance industry won't tell you: the Supreme Court's decision in SEC v. Jarkesy creates a serious question about whether the IRS can impose these penalties without giving you the right to a jury trial. Jarkesy held that civil penalties that are punitive in character — that look and function like common law fraud penalties — trigger the Seventh Amendment right to jury trial.

The erroneous refund claim penalty and accuracy-related penalty in a Section 41 case are potentially punitive in character. If the IRS imposes them, you may have a constitutional right to have a jury of your peers decide whether you owe them. That is not a right the IRS will volunteer to explain to you.

Who Is Actually at Risk — And Who Has a Stronger Position Than They Think

The IRS focuses its Section 41 enforcement resources on specific targets. Large corporations — those with assets over $10 million — actually have a structured safe harbor pathway through the LB&I Directive that can get their credit accepted. The sophisticated companies with CFOs and tax departments have options.

The real targets are small and mid-size businesses that were sold credit studies by contingency-fee promoters. These taxpayers received prepackaged documentation that described a methodology but often failed to actually substantiate that specific qualified research expenses were incurred. When the IRS examines the return, they find thick binders and thin substance.

But here's the Reality Check: many of these taxpayers have a stronger position than they or their current advisors realize.

If the company genuinely undertook investigation to resolve uncertainty — even if the credit study overreached on the amount — there is real research credit there. The question is whether anyone is willing to go back to the underlying facts and build the correct argument rather than either surrendering to the IRS or defending an indefensible position.

The Loper Bright argument — that the IRS's regulatory definition of qualified research is narrower than the statute permits — is available to every taxpayer whose credit was disallowed because their research produced dead ends rather than successful technological development.

The Jarkesy argument — that civil penalties require jury trial — is available to every taxpayer facing substantial penalties.

Neither of these arguments will come from your CPA. These are arguments that require someone willing to challenge the IRS's legal framework rather than accept it as correct.

The New Form 6765 Trap

Beginning with tax year 2024, the IRS revised Form 6765 to require detailed business component information — descriptions of the uncertainty investigated, the alternatives evaluated, the type of business component, qualifying wages by category, and more.

Many taxpayers who claimed research credits in prior years based on promoter-prepared studies have no documentation that meets these new requirements. If their return is selected for examination, they cannot produce the information the IRS now demands.

This is a documentation problem, not necessarily a substantive problem. If the underlying research activities genuinely qualified, the solution is rebuilding the documentation correctly — not surrendering the credit.

The OBBBA Opportunity You May Be Missing

The One Big Beautiful Bill Act, signed July 4, 2025, made immediate expensing of domestic research expenses permanent under new Section 174A and created a retroactive election opportunity for small businesses.

If your company had average annual gross receipts under $31 million, you may be able to retroactively elect immediate expensing of research expenses for tax years 2022 through 2024. The deadline is July 6, 2026 or the applicable statute of limitations, whichever is earlier.

This matters for Section 41 purposes because the interaction between research expense deductions and the research credit — governed by Section 280C — can significantly change the economics of your position. If you're currently in examination on a research credit claim, understanding the OBBBA's impact on your overall tax position is essential before you settle.

What IRSMedic Does Differently

The compliance industry has one approach to a Section 41 examination: accept the IRS's regulatory framework as correct, try to fit your facts into the IRS's definition of qualified research, and negotiate the disallowance down as far as possible.

We don't start from the assumption that the IRS's interpretation of the statute is correct.

Post-Loper Bright, the IRS's regulatory definition of qualified research is not entitled to judicial deference. Courts must independently determine what "research activities" means in the statute. We believe the correct answer — the one the statute actually supports — is broader than the IRS's regulations permit.

If you're facing a Section 41 examination, you need someone who will investigate your specific situation, identify what genuine research activities occurred, build the documentation that supports those activities, and challenge the IRS's legal framework where the framework itself is wrong.

That's what IRSMedic does. We are advocates for our clients — not compliance officers for the IRS.

If a third-party firm sold you an R&D credit study and the IRS is now examining your return, contact us before you respond to any IRS correspondence. The position you take in the examination phase shapes everything that follows.

IRSMedic.com | IRSMedicMembers.com

For Top Level Help With a Tax Problem book a call with us at : IRSMedic.com Join our gorwing community at IRSMedicMembers.com