How Custom Manufacturers Can Claim the Federal R&D Tax Credit

TL;DR

• Custom manufacturers frequently solve technical problems on the floor (new tooling, new materials, tighter tolerances) that may meet the IRS definition of qualified research.
• The federal credit under IRC Section 41 turns on a four-part test, not on having a lab or a formal R&D department.
• A no-cost eligibility assessment can tell you whether your shop's work likely qualifies before you commit to anything.

You solved it on the floor. That might be R&D.

Most custom manufacturers do not think of themselves as doing research. You think of it as doing the job: a customer brings a part nobody has made quite this way before, and your team figures out how to hold the tolerance, pick the material, build the fixture, and get a repeatable process running. That figuring-out is exactly the kind of activity the federal R&D tax credit was written to reward, and a lot of shops in the $3M to $15M range leave it on the table because the work never carried the label "research."

The credit is not reserved for white coats and clean rooms. It can apply to a machine shop, a fabricator, a mold maker, or a contract manufacturer that is regularly developing or improving products and processes. The question is not whether you have an R&D department. It is whether specific activities meet the test the IRS uses.

What the federal R&D credit actually covers

The credit comes from Internal Revenue Code Section 41, and the supporting rules live in Treasury Regulation 1.41-4 (read the regulation here). To be treated as qualified research, an activity generally has to clear all four parts of what the IRS calls the four-part test:

Technological in nature. The work has to rely on principles of a hard science such as engineering, physics, chemistry, or computer science. For a custom manufacturer, machining, materials science, and mechanical engineering typically fit this part comfortably.

Permitted purpose. The activity has to aim at creating a new or improved product or process, measured by function, performance, reliability, or quality. Developing a part that holds a tighter tolerance, or a process that runs with less scrap, can satisfy this.

Technical uncertainty. At the outset, you could not be certain of the capability, the method, or the appropriate design. If you knew exactly how the job would run before you started, that points away from qualification. If you had to test your way to the answer, that points toward it.

Process of experimentation. You had to evaluate alternatives in a systematic way: modeling, simulation, trial runs, first-article testing, or iterating on tooling until it performed. This is the part that good documentation tends to make or break.

The IRS guidance on the credit is summarized on IRS.gov. Whether any given project qualifies depends on the specific activities, the documentation, and the facts, so the right read for your shop comes from looking at actual jobs, not at the industry in general.

What qualifying activity often looks like in a custom shop

In custom manufacturing, the qualifying work is frequently hiding inside ordinary quoting and production. Activities that often qualify, depending on the facts, include: designing and proving out new tooling, jigs, or fixtures for a part that has not been run before; developing a process for a new or difficult material; engineering to hit a tolerance or surface finish your current process cannot reliably hold; building and testing prototypes or first articles ahead of a production run; and reworking a process to reduce cycle time, scrap, or rejects where the path to get there was not obvious.

Documentation is what carries these from "we just did it" to "here is the qualified research." Time records tied to projects, engineering notes, revision histories on programs and drawings, scrap and yield data, and first-article inspection reports all help substantiate both the uncertainty you faced and the experimentation you ran.

What usually does not qualify

Being honest about the boundaries matters as much as finding the credits, because it is what keeps a claim defensible. Work that typically does NOT qualify includes: routine production of a part you already know how to make; standard quoting and estimating; ordinary quality control and inspection of finished goods; cosmetic or styling changes with no functional or performance objective; and simply buying and installing a machine that works as the manufacturer intended, with no development of your own process around it.

Research conducted outside the United States, and work funded by a customer in a way that shifts the financial risk and rights away from you, also generally falls outside the credit. These funding and ownership questions are fact-specific and are worth reviewing job by job rather than assuming either way.

What to do next if you are an owner or CFO

If your shop regularly takes on parts or processes that have not been run before, there is a reasonable chance some of that work qualifies, and it is worth a look. A typical next step is a no-cost initial assessment: a specialist reviews the kinds of projects you run and the records you keep, and tells you whether a credit study is likely worth pursuing before you spend anything. From there, a completed study commonly runs in the range of four to eight weeks, depending on the size of the shop and how readily the documentation comes together.

Strata's focus is on capturing the credit you have actually earned and documenting it so it holds up, rather than stretching a claim past what the facts support. If you want to find out where your shop stands, start with a no-cost assessment.

FAQ

We do custom one-off jobs, not mass production. Can we still qualify?
Possibly, and in some shops the one-off and short-run work is where the strongest qualifying activity lives, because each new part can carry its own technical uncertainty. Qualification still depends on meeting the four-part test for the specific activities involved.

Do we need a formal R&D department or written research plans?
No. The credit looks at what the work involved, not at your org chart. That said, contemporaneous records (project time, engineering notes, test and inspection data) make a claim far easier to support, so the absence of formal plans is not a barrier but good documentation is a real advantage.

Does buying a new CNC machine count as R&D?
Buying equipment that performs as intended generally does not, on its own. Developing a new process around that equipment, or solving technical problems to run a part the machine could not handle out of the box, is where qualifying activity may arise, depending on the facts.

What if our customer paid us to develop the part?
That depends on the terms. Where a customer funds the work and holds the financial risk and rights to the results, the credit may not be available to you. Because funded-research rules are fact-specific, contracts are usually worth reviewing case by case before drawing a conclusion.

This post is for informational purposes only and does not constitute tax or legal advice. Consult a qualified tax professional regarding your specific circumstances.