Manufacturers Are Leaving R&D Credits on the Table

Ask a manufacturer whether they do R&D and you will often hear “no - we don’t have a lab, we just make parts.” That answer leaves money on the table. The federal R&D tax credit was written to reward the kind of technical problem-solving that happens on a shop floor every day: figuring out how to make a part faster, with less scrap, on new equipment, or out of a material you have never run before. The credit is not reserved for white-coat research. For many manufacturers, the most valuable qualifying work is the engineering they already do to keep production moving.

The “you need a lab” myth

The R&D credit under IRC Section 41 does not require a laboratory, a research department, or a breakthrough invention. It applies to activities that meet a four-part test:

1. Permitted purpose - improving the functionality, performance, reliability, or quality of a product or a process.
2. Technological in nature - relying on engineering or the physical sciences.
3. Elimination of uncertainty - you did not know, at the start, whether or how you could achieve the result.
4. Process of experimentation - you worked through the uncertainty by evaluating alternatives, testing, and iterating.

Notice the words “or a process.” This is the part manufacturers most often miss. You do not have to be designing a new product to qualify. Developing or improving the process by which you make something can qualify on its own. That single point opens the credit up to a large share of ordinary manufacturing engineering. See our manufacturing industry page for how this plays out across different shops.

Process development and shop-floor engineering

Process development is the most overlooked category of qualifying activity in manufacturing precisely because it does not feel like “research” - it feels like the job. When your team faces a production problem with no known solution and works through it with engineering and testing, that is a process of experimentation aimed at improving a process. Common examples:

• Tooling and fixture design - developing new tooling, dies, molds, or fixtures where the right design is uncertain and has to be tested and refined.
• Automation and robotics integration - designing, programming, and tuning automated cells, robotics, or material handling where you had to figure out whether and how it would work in your environment.
• Scrap and yield improvement - re-engineering a process to reduce defects, scrap, or rework, when the path to a better yield was not known in advance.
• New materials - qualifying and dialing in a process for a material you have not run before, where parameters had to be discovered through trials.
• First-article and pilot runs - developing and validating a new manufacturing process, including the trial runs needed to prove it can produce conforming parts at the required tolerances.
• CNC programming and process parameters - establishing feeds, speeds, sequences, and parameters for a new or difficult part through testing rather than from a known recipe.

The connecting thread is uncertainty resolved through experimentation. If the answer was already in a handbook or a vendor’s spec sheet and you simply followed it, it likely does not qualify. If your engineers and machinists had to run trials to find out what would work, it likely does. The four-part test in plain English walks through each part and where shop-floor work passes or fails.

The exclusion that catches contract shops: funded research

Before you total anything up, screen your contracts. Under §41(d)(4)(H), research is excluded to the extent it is funded by a grant, a contract, or another person - and the test turns on two questions that live in the purchase order, not in the engineering:

• Do you retain substantial rights in the results? If the customer takes exclusive ownership of the process, tooling design, or data you developed, your claim is in trouble.
• Do you bear the economic risk of failure? If you get paid for the development effort whether or not it works - cost-plus, time-and-materials, or a reimbursed NRE line - the risk sat with the customer.

For a contract manufacturer, the classic pattern is customer-funded NRE and tooling development: the customer pays for the tool, owns the tool, and pays for the engineering hours regardless of outcome. That work can be excluded no matter how genuinely uncertain it was. The mirror-image pattern - fixed-price work where you eat the overruns and keep the process know-how - often survives.

This is not a reason to skip the credit. It is a reason to sort your jobs by contract type before you compute anything, and to keep the contracts in the study file. Our manufacturing page goes deeper on how this plays out across job shops and OEM suppliers.

Which costs count (QREs)

For activities that qualify, the credit is computed on qualified research expenses (QREs):

• Wages for qualified services - compensation for the people performing, supervising, or directly supporting the qualifying work. For manufacturers, this often includes process and manufacturing engineers, and the machinists and technicians involved in developing and testing the new process - not just a dedicated R&D staff.
• Supplies consumed in the research - and this is a category manufacturers should pay close attention to. The material consumed during trial runs, scrap produced while developing a process, and prototype or first-article parts can be qualifying supplies. When you burn through raw material proving out a new tooling setup, that consumed material may count.
• Computer and cloud rental - costs of computers used in qualifying research, such as simulation or modeling work, recognized under Section 41(b)(2)(A)(iii).
• Contract research - generally 65% of amounts paid to another person for qualified research performed on your behalf (§41(b)(3)(A)), such as an outside engineering firm helping develop a new process. The research has to be performed in the United States (§41(d)(4)(F) excludes research conducted abroad), and you have to retain substantial rights and bear the economic risk - otherwise it is funded research to you.

Trial and scrap supplies deserve a second mention because they are so frequently left out. A development effort that consumes significant material in iterations is generating QREs in those iterations, but only if someone tracks them. One boundary to respect: land, and property subject to depreciation, are not supplies - the production tooling you capitalize is not a supply QRE, even though the trial material consumed proving it out can be. Our qualified research expenses page draws the full line.

Documentation matters

Process development qualifies, but qualifying and proving it are two different things. Because shop-floor experimentation lives in travelers, run sheets, engineering change orders, scrap logs, and machinists’ notes rather than in a formal research report, the documentation is often scattered or thin. A solid study connects the qualifying activities to the four-part test using contemporaneous records - the kind of evidence created while the work is happening, not reconstructed years later.

The revised Form 6765 reinforces this with Section G, which calls for business-component-level detail about the research. Per the current Instructions for Form 6765, Section G is optional for tax years beginning before 2026 and applicable for tax years beginning after 2025, with exceptions for qualified small businesses electing the payroll credit and for smaller filers claiming on an originally filed return. Those dates and thresholds have already moved more than once - confirm against the current-year instructions.

In practice it means being able to show, component by component, what you were trying to improve, what was uncertain, and how you experimented. Manufacturers that capture this as they go are in a far stronger position than those scrambling at filing time. What examiners actually ask for covers the evidence chain in detail.

The takeaway

If your team routinely solves production problems that do not have a known answer - new tooling, tighter tolerances, less scrap, a material you have not run, a line you are automating - you are very likely doing qualifying research, lab or no lab. The opportunity for manufacturers is threefold: recognize that process development counts, track the trial and scrap supplies and the engineering time that come with it, and sort your jobs by who funded them before you compute anything. To see how the credit fundamentals apply to your operation, start with our R&D tax credit overview, the four-part test, and state R&D credits - many states run a piggyback credit off the same qualified base.

Sources

• IRC §41 - Credit for increasing research activities (U.S. House, Office of the Law Revision Counsel)
• Treas. Reg. §1.41-4 - Qualified research (Cornell LII)
• IRS - Research credit
• IRS - About Form 6765 and the Instructions for Form 6765